Expanding the clinical-pathological and genetic spectrum of RYR1-related congenital myopathies with cores and minicores: an Italian population study

Mutations in the RYR1 gene, encoding ryanodine receptor 1 (RyR1), are a well-known cause of Central Core Disease (CCD) and Multi-minicore Disease (MmD). We screened a cohort of 153 patients carrying an histopathological diagnosis of core myopathy (cores and minicores) for RYR1 mutation. At least one RYR1 mutation was identified in 69 of them and these patients were further studied. Clinical and histopathological features were collected. Clinical phenotype was highly heterogeneous ranging from asymptomatic or paucisymptomatic hyperCKemia to severe muscle weakness and skeletal deformity with loss of ambulation. Sixty-eight RYR1 mutations, generally missense, were identified, of which 16 were novel. The combined analysis of the clinical presentation, disease progression and the structural bioinformatic analyses of RYR1 allowed to associate some phenotypes to mutations in specific domains. In addition, this study highlighted the structural bioinformatics potential in the prediction of the pathogenicity of RYR1 mutations. Further improvement in the comprehension of genotype–phenotype relationship of core myopathies can be expected in the next future: the actual lack of the human RyR1 crystal structure paired with the presence of large intrinsically disordered regions in RyR1, and the frequent presence of more than one RYR1 mutation in core myopathy patients, require designing novel investigation strategies to completely address RyR1 mutation effect.


Introduction
Core myopathy are clinically and genetically heterogenous congenital myopathies defined by the presence on skeletal muscle biopsy of "cores" with focally reduced oxidative activity, lack of mitochondria and variable degrees of myofibrillar disruption [1]. Based on the histological findings on muscle biopsy, core myopathies were traditionally classified in Central Core Disease (CCD), where the cores, in transverse sections of the myofibers, are single in the centre of the fibre and span through all the length of type 1 muscle fibre, and Multi-minicore Disease (MmD) with multiple and less defined cores without specificity for fibre type [2,3]. This binary classification of core myopathies is becoming reductive since in the same muscle biopsy both CC and Mm may exist, a transition of the main histopathology features [4] or the presence with cores of other structural abnormalities in muscle fibres (i.e. rods [5,6]) is common. To add to this complexity, new entities are entering in the picture. The Dusty Core Disease (DuCD) is quite recent, and it is defined by irregular areas of myofibrillar disorganization with reddishpurple granular material depositions that show uneven oxidative staining and are devoid of ATPase activity [7]. Dusty cores, in contrast with central cores, have no clear borders and are not round/ovoidal in shape [7].
CCD is clinically characterized by hypotonia, motor developmental delay and generalized muscle weakness of variable degree. Distribution of weakness is typically proximal, with prominent involvement of the hip girdle and axial muscles. Bone and joint disorders, such as congenital hip dislocation, kyphoscoliosis, pes cavus, thoracic deformities, joint contractures are often present. The age of onset is extremely variable, ranging from foetal akinesia-which has been primarily reported associated with recessive inheritance [9], or birth onset-with hypotonia and floppy infant syndrome, childhood onset with motor developmental delay, but also adult onset with variable clinical presentation, often within the same family [3]. The penetrance of the disease is variable and genotype-phenotype correlations have yet to be fully clarified [10]. The association with malignant hyperthermia (MH), a life-threatening pharmaco-genetic disorder triggered by exposure of susceptible individuals to inhalational anesthetics and succinylcholine, is strong, though variable, and all patients suffering from core myopathies should be considered susceptible [11].
Both CCD and MH are primarily caused by mutations in RYR1, which encodes ryanodine receptor 1 (RyR1). RyR1, located on the sarcoplasmic reticulum membrane, is a Ca 2+ channel comprising of four identical subunits [12], each one of 5,038 amino acids, with a molecular weight of 563.5 kDa [13]. The monomers delimit a central ion-conducting pore. RyR1, interacts with the dihydropyridine receptor (DHPR), and it is involved in the excitation contraction coupling, that ensures skeletal muscle contraction upon stimulation [14,15].
The RyR1 protein is made of a cytoplasmic shell (also known as "foot") and domains constituting the RyR1 channel and the activation core. The foot comprises: the N-terminal domains (NTD-A, NTD-B, NTD-C), the bridge solenoid (BSol), the junctional solenoid (JSol)which connects NTD-C to BSol, the SP1a/ryanodine receptor domains (SPRY1-SPRY3) and the RyR repeats pairs (RY1&2, RY3&4). The RyR1 channel portion and the activation core comprise several structural domains: the shell-core linker peptide, CaM, and JSol binding sites (SCLP), the core solenoid (CSol), the thumb and forefingers domain (TaF), the auxiliary transmembrane helices (TMx), the pseudo voltage sensor domain (pVSD), the channel pore domain, made of six transmembrane helices, and the C-terminal domain (CTD) [16]. This region is directly responsible for the Ca 2+ -mediated channel activation [12,17]. Moreover, given its crucial role in muscle physiology, RyR1 activity is finely regulated through a wide variety of post-translational modifications of the channel, including oxidation, phosphorylation [18], and the binding of ligands. In particular, calmodulin (CaM), calstabin1 (previously known as FK506-binding protein 12, FKBP12) and CLIC2 [19] interact with the CTD, and triadin, junctin, and calsequestrin with the cytoplasmic shell [12,19].
Autosomal dominant RYR1 mutations are identified in approximately 90% of CCD and are mostly localized in three hotspots: the cytoplasmic N-terminus (hotspot 1; amino acid 35-614), the central domain (hotspot 2; amino acid 2163-2458) and the C-terminus (hotspot 3; amino acid 4550-4940) [3,20]. On the basis of functional studies, distinct molecular mechanisms were proposed to explain how specific RYR1 mutations could result in core myopathies and/or in MH [21]. The clinical spectrum of MmD is also variable and depends on the genetic background [22]. MmD patients may manifest hypotonia and proximal muscle weakness, ophthalmoparesis, atrophy of the small muscles of the hands. Severe neonatal-onset form have been identified [22][23][24][25].

Patients
An Italian network of tertiary referral Centers for congenital myopathies had been established in 2008 to collect detailed clinical, morphological, and genetic data in a large group of Italian patients. One hundred fifty-three patients were selected according to muscle histopathology consistent with cores or minicores myopathy and chosen for RYR1 mutations analyses. Ethics committees of all participating centers approved the study and written informed consent was obtained from all patients or their legal guardians, in accordance with the ethical standards of the 1975 Declaration of Helsinki, revised in 2013. Clinical evaluation was performed at each referring Center according to a standardized protocol, including the collection of family history, disease onset and progression data, functional abilities, muscle strength and severity of contractures when present. Electrocardiogram, echocardiogram, pulmonary function test results were collected for each patient. Based on disease severity, patients were classified as: "asymptomatic" if they came to medical attention for CK elevation, without symptoms of neuromuscular disease; "paucisymptomatic" if they complained of myalgias, and/or muscle cramps but not overt muscle weakness. In this subset of patients, the presence of minor osteoarticular alterations was allowed; "myopathic" if muscle weakness was present. This last subset was further categorized according to the MRC score of proximal and distal muscles in the upper and lower limbs in mild (MRC score ≥ 4 in all muscle tested), moderate (MRC ≤ 3 in one proximal muscle) and severe (MRC ≤ 3 in two or more proximal muscles). The age of onset in symptomatic patients was defined as "congenital" if clinical symptoms appeared before 12 months of age; "early" if symptoms were present before age of 20; "adult" if symptoms appeared after 20 years of age.

Muscle biopsy analysis
Muscle biopsies, done at the time of diagnosis, were reviewed, when available, at each referring centre by an expert myopathologist. Transverse cryosections of muscle biopsies stained with NADH-TR (nicotinamide adenine dinucleotide dehydrogenase-tetrazolium reductase), COX (cytochrome C oxidase) or SDH (succinate dehydrogenase), haematoxylin-eosin or Gomori's trichrome were examined on upright microscope (Olympus BX60, Tokyo, Japan). All fibres with central nuclei, cores in type 1 or type 2 muscle fibres, rods, and the number of type 1 fibres were counted in 3 independent microscopic fields (at a magnification of 10× ). The results had been expressed as mean percentage.

RYR1 mutations analysis
Genomic DNA was extracted from peripheral blood according to standard procedure. The entire coding sequence of RYR1 (NG_008866.1) was amplified from patient genomic DNA and analysed by Sanger sequencing. Frequency of mutation in the hotspots was assessed with proportion test performed with R v. 3.5.3.

Homology modelling of the human RyR1 3D structure
The human RyR1 sequence (accession code: P21817) was obtained from Uniprot [36] and used to perform a Blastp search against the Protein Data Bank [37]. The 3D structure of the RyR1 orthologous protein from Oryctolagus cuniculus (PDB ID 5T15; electron microscopy structure, resolved at 3.6 Å) was selected as a template to build the model based on its high sequence identity and coverage with the human RyR1 (identities = 96.91%, coverage = 72%). As this structure presents multiple not resolved segments, the human RyR1 sequence was aligned against the template with Jalview [38] using T-Coffee [39] algorithm (default parameters) to identify the protein domains. Six comparative models of the target sequence were built by MODELLER [40] using the model single module. Models were then evaluated for value of discrete optimized protein energy using the DOPE method integrated in MODELLER. Model number 2 was selected as the best final model. Overall structure quality was further assessed with QME-ANDisCo [41] and MolProbity [42]. The MobiDB-lite [43] was used to exactly identify and map intrinsically disordered regions boundaries mostly overlapping segments not resolved in the template structure. The model presents a good geometry with 90.09% of residues in the most Ramachandran favoured regions while the 2.05% are marked as outliers. The Rama distribution Z-score calculated for the entire model is -0.98 ± 0.11 (Goal: < 2) whereas the number of bad bonds is estimated to 0.01% of total. To avoid introduction of unwanted artifacts there was no attempt made to manually adjust the backbone torsion angles. The final 3D model corresponds to the human RyR1 segments spanning residues 24-1272, 1432-1924, 2057-2562, 2734-2940, 3645-4250, 4546-5034.

Bioinformatics analyses of RYR1 mutations
An integrative in silico pipeline was used to evaluate pathogenic impact of RYR1 mutations described in Additional file 1: Fig. S1. In details, a total of 45 RyR1 homologous sequences were retrieved by PSI-BLAST [44] search (default parameters) against the UniProt sequence database and aligned with T-Coffe (default parameters) [39] to derive a conservation score used each residue. This conservation score was then used as a preliminary filter to evaluate impact of substitutions found in patients. The ELM [45] database was used to search and map functional linear motifs, while FELLS [46] was used to derive structure features. Structural inspection was performed with Chimera [47], while effect of mutations on the human RyR1 3D structure was evaluated with BLU-UES [48] and RING2.0 [49]. Stability and pathogenicity assessment were carried out using a consensus approach including predictions from Polyphen2.0 [50], SIFT [51], Mupro [52], Mutationassessor [53] (Additional file 1: Fig. S1). The new Intronic variant has been analysed with two independent algorithms for splice signal detection: NetGene2 (http:// www. cbs. dtu. dk/ servi ces/ NetGe ne2/) and Berkeley Drosophila Genome Project (BDGP, https:// www. fruit fly. org/ seq_ tools/ splice. html).

Statistical analyses
In order to evaluate associations between the presence of mutations in specific domains and several phenotypic traits (foetal hypokinesia, hypotonia at birth, arthrogryposis, respiratory insufficiency, dysphagia, delayed independent ambulation, congenital hip dislocation, facial weakness, muscle wasting, ocular involvement, bulbar involvement, contractures, foot deformities, scoliosis/rigid spine, cardiac abnormalities, cognitive issues), for each pair of protein domain and phenotypical trait, a 2 × 2 table was devised, with rows indicating presence vs. absence of mutations in that domain, and columns indicating presence vs. absence of the phenotypic trait. P-values < 0.05 are reported (without correction for multiple testing) and should be interpreted as descriptive.

Results
At least one RYR1 mutation was identified in 69 core myopathy patients and these have been further studied (Table 1).

RYR1 mutations
Sixty-eight different nucleotide variations in RYR1 sequence were identified in 69 patients from 55 unrelated families. As expected, mutations were localized with higher frequency in the 3 hotspot regions, than in the rest of the sequence (p < 0.0001). Variations included 59 missense mutations, 3 splice site variants, 4 small frameshift insertion or deletion, and two in-frame deletion of a single amino acid (Table 1). Among the 68 mutations, 16 were novel (Table 1). Twenty-three patients carried more than one RYR1 variant (Table 1); of these patients, 13 (57%) were isolated case, 2 (9%) patients had an autosomal dominant (AD) and 8 (35%) an autosomal recessive (AR) inheritance pattern. Unfortunately, parents of patients were not available for analyses and the phase of the mutations was not assessed, with the exception of patients #17 and #69 who carry mutations known to be transmitted in cis [77,90].

Muscle histopathology
Muscle biopsy, done at time of diagnosis, was reviewed in 52/69 patients. Of the remaining 17 patients, 10 were familial cases where only a relative had the muscle biopsy done and in 7 patients muscle biopsy, consistent with core myopathy according to the referring physician, was not available for review. Quadriceps was biopsied in 45 patients, deltoid in 3, biceps and triceps in one each, while there was no information about the site of biopsy for 2 subjects. For all the samples, only transverse sections were available and analysed. All biopsies but one (#31, Additional file 3: T2) showed cores in a variable number of type 1 muscle fibres (ranging from 8 to 100%). Patient #31 had cores in 39% of type 2 muscle fibres only. In type 1 fibres, cores were centrally located (central cores, CC) in 67% of muscle biopsies, multiple minicores (Mm) and CC were present in 15% and Mm alone in 17% of muscle biopsies. CC and Mm were mostly observed in type 1 muscle fibres, whereas a minority of patients showed CC and Mm in type 2 fibres. No dusty cores were observed in this cohort of patients ( Fig. 1).
Type 1 fibre predominance (defined as more than 45% of type 1 fibres) [91], ranging between 50-100%, was observed in 44 muscle biopsies (44/50; 88%) and increase in central nuclei (> 3% of fibres) was detected in 33 (33/50; 66%). An increase in perimysial and endomysial connective tissue was mild in 24 (24/51; 47%), severe in 6 (6/51; 12%) and absent in 20 (20/51;   to verify if mutations located in specific RyR1 domains were associated with more severe phenotype, we grouped patients based on the domain affected by the RYR1 mutation and utilized clinical data at last evaluation for comparison. Subjects carrying more than one pathogenic mutation were described in each mutated domain group ( Fig. 2 and Additional file 2: T1). Variations predicted or known to be benign were excluded from the analysis. Frame-shift mutations have been studied as separate group, independently from the affected domain. Results are summarized in Fig. 2 and Additional file 2: T1, and detailed clinical data are reported in Additional file 3: T2.
• NTD-A domain mutations. In our cohort 7 patients had mutations in NTD-A domain, only one of them (patient #47) had a severe clinical course. She presented at birth as a floppy infant, independent ambulation was acquired at 16 months and at 12 years of age the patient had a severe proximal and distal muscle weakness. Patient #1 had a congenital onset with hip dislocation, but an overall mild clinical course and at age 16 years the patient has a mild myopathy and retains the ability to walk and run. Muscle weakness was reported by 57% of patients but disease progression was stationary in all but one. The association between mutations in NTD-A with facial weakness and delayed independent ambulation was nominally significant (Fisher exact test p = 0.039, p = 0.038, respectively). • NTD-B domain mutations. Two patients had mutations in NTD-B domain. Both patients presented at birth with arthrogryposis. Patient #7 had a severe AR myopathy with proximal and distal muscle weakness but was last seen at 3 years of age, while patient #8, who had delayed independent ambulation at 24 months of age had a subsequent slowly progressive disease and at age 57 years she is still ambulant with a moderate myopathy.  (7), one patient was identified through family screening and was asymptomatic at 1 < 6 years of age and for 2 patients no information were available regarding disease onset. Eight had a mild, 5 moderate and 3 a severe myopathy. Four patients showed facial mus-  (2). Muscle weakness was moderate in 2, mild in 4 and one patient was paucisymptomatic. Two patients had a moderate myopathy, and one of them had also rhinolalia and dysphagia.

RYR1 mutations bioinformatic analyses
Four RyR1 chains assemble to form a single calcium release channel and multiple 3D structures derived from cryomicroscopy studies describing their organization are available in the literature. To investigate the effect of mutations in our dataset, we generated a homology model of the human RyR1 monomer using the orthologous 3D structure from rabbit (identities = 96.91%, coverage = 72%). Resulting model presents an overall good geometry with 90.09% of residues occupying the Ramachandran favoured regions. The 2.05% of residues are estimated to occupy unfavourite positions. Visual inspection shown that most of them localize at the borders of regions predicted as intrinsically disordered. Pairing these data with a number of bad bonds corresponding to 0.01%, we suggest that our model is good enough to assess the effect of mutations found in patients.
In details, mutations were classified as "pathogenetic from literature" if already reported so, "pathogenetic from function" based on in silico studies, "likely benign" based on literature data and/or bioinformatics analyses, and "truncating, frameshift and intronic" based on their effect on the protein sequence.

RYR1 mutations "pathogenetic from literature"
Thirty-eight mutations were identified in this subset.

"Pathogenetic from function" RYR1 mutations
We modelled fifteen variants, which were either novel, or not previously studied (11 variants) or with unclear effect on the RyR1 channel from literature (4 variants) (i.e. p.Val1101Met [65], p.Gly3938Asp [89], p.Arg4564Gln [7], p.Arg1606Cys [70]). The human RyR1 (Uniprot code: P21817) was analysed with a collection of different bioinformatics tools to evaluate the impact of amino acid mutations on the RyR1 protein function.
• NTD-A domain. The mutation p.Glu158del localizes in a flexible loop connecting two β-strands at the beginning of the MIR2 domain. Pathogenicity predictors suggest that this mutation reduces the domain stability. Further, multiple mutations in this area are already known to promote both MH and CCD [92]. • RY1&2 domain. The mutation p.Thr977Met was predicted as damaging. Thr977 is exposed to the solvent and structural investigation suggests its variation in methionine to destabilize the correct RY1&2 (amino acids 850-1054) domain folding. Previous observations proposed that this repeated region, together with the SPRY domains, acts as calstabin1 interacting region [103]. A pathological phenotype 1n patients harbouring the p.Thr977Met may be associated with impairment of this interaction.
• SPRY2/SPRY3 domain. The p.Val1101Met is predicted as probably damaging as Val1101 engages hydrophobic interactions with Phe1089 and Phe1091 stabilizing the SPRY2 domain spanning residues 1055-1241. According to literature, the region within residues 1272-1455 is responsible for interaction with the II-III loop of DHPR [94]. Experimental validation demonstrated that the region is relevant for skeletal muscle contraction in vivo [94], suggesting that variants p.Arg1289Gln and p.Arg1301Cys may interfere with the DHPR-mediated regulation of RyR1. A damaging effect on the protein structure is predicted for p.Arg1606Cys. Structural inspection shows Arg1606 to form electrostatic interactions with both Asp1111 and Glu1113. At the structural level, our analysis suggests that this mutation alter correct localization of the SPRY2 domain inducing local protein unfolding. Damaging effect is also predicted for p.Leu1650Pro. This position is well conserved among eukaryotes indicating it may play a relevant structural and/or functional role. Our structural investigation showed Leu1650 to form multiple van der Waals interactions with residues from the adjacent chain, suggesting that its mutation can negatively affect RyR1 tetramerization, yielding a functionally reduced channel. Mutation p.Gln1837Pro localizes in a short alpha helix segment connecting two longer helices spanning residues 1803-1852. The entire segment is part of a larger junctional solenoid domain. As proline is known to destabilize secondary structure, we believe the mutation induces a local unfolding of this domain. • BSol and CSol domains. Mutations p.Leu3049Phe, p.Lys3098Ile, p.Gln3506Lys and p.Thr3513Ala localize in a long disordered region important for RyR1 association with CaM (Fig. 2). All of them are predicted as damaging, possibly interfering with CaM mediated RyR1 regulation. A different pathogenic effect is predicted for p.Gly3938Asp. The Gly3938 localizes at the internal binding interface between RyR1 monomers directly facing the residue Glu79 from an adjacent monomer chain. The repulsive interaction introduced by p.Gly3938Asp is rather predicted to lower RyR1 complex stability than having a negative effect on monomer structure. • pVSD, Pore and CTD domains. The region between residues 4300-4850 is known to be an important regulative element for the RyR1 activation/inactivation switch [98-100], with several known pathological mutations falling within. In particular, p.Arg4564Gln impairs an electrostatic interaction between Arg4564 and Tyr4792. The analysis of mutations localizing into the interval 4505-5008, i.e. p.Asp4939Tyr and p.Glu5008Gln, predict them as damaging and suggest a potentially pathological phenotype, affecting the apoCAM binding domain functionality and the ions flux balance regulation. The evidence is also in agreement with [104].

"Likely benign" RYR1 mutations
• NTD-A domain. In silico pathogenicity prediction suggests p.Tyr564His to be tolerated. A modest structural impact is also predicted by investigating the residue interacting network around Tyr564 (Fig. 3). Indeed, investigation shows p.Tyr564His to localize in the N-terminal solenoid spanning residues 393-627. Based on our model, this residue is partially exposed and modestly involved in the tandem repeat domain folding. We thus suggest this mutation to be likely neutral. • SPRY2/SPRY3 domain. As for p.Tyr564His, sequence conservation analysis shows that Pro1312 is only limitedly conserved among different species, reinforcing the idea that mutation p.Pro1312Leu may be tolerated. Another possible explanation of this result is that stability predictors fail to address p.Pro1312Leu pathogenicity as Pro1312 localizes in an intrinsically disordered region. The p.Ile1571Val [62,67] variant is reported to be benign. • JSol domain and intronic region. p.Pro1787Leu [56], p.Gly2060Cys [54,56,105], and c.10259 + 7G > A are reported to be benign variants.

Truncating and frameshift RYR1 mutations
We found four frameshift mutations i.e. p.Pro2361AlafsX2, p.Leu3136Argfs, p.Gln4837ArgfsX3 and p.Asp4021GlufsX4. In our series, a predicted altered quantitative level of RyR1 is consistent with stable or slowly progressive mild/moderate myopathy, suggesting that the level of RyR1 protein may be critical for the normal excitation-contraction coupling [56,60]. The c.10347 + 1G > A mutation had already been described as probably damaging. The mutation likely abolishes the donor splice site, producing a shorter or unstable mRNA and it was predictably damaging by in silico analyses [88]. Finally, we found also a new intronic variant i.e. c.11609-2A > G, localized in a highly conserved region. Different bioinformatic tools predicted that the mutation causes the loss of a splicing-acceptor site in intron 83. The mutation likely leads to the production of an aberrant transcript with the partial out-of-frame retention of intron 84.

Discussion
In this paper we collected clinical, histological and molecular data from a large cohort of core myopathies patients carrying RYR1 mutations with the intent to use in silico modelling to evaluate the possible pathogenetic impact of the identified RYR1 variants, and to verify if variants targeting to specific RyR1 domain are associated to a more severe phenotype.
We used a candidate gene approach using RYR1 Sanger sequencing to screen a large cohort of core myopathy patients. One hundred fifty-three patients were enrolled and RYR1 mutations were identified in 69 (45%) of them. The lack of detailed histopathological information in most of the original cohort of patients, did not allow to assess the overall prevalence of RYR1 mutations in core and multi-minicore diseases. Nevertheless, since CCD is reported to be mainly due to RYR1 mutation [20], while MmD are associated with mutations in variety of genes (RYR1, but also SEPN1, and less frequently in MYH2, UNC45B, MYH7, TTN, MEGF10, SECISBP2, ACTA1, ACTN2, CCD78, and FXR1) [8] we can hypothesize in our original cohort a majority of MmD patients carrying mutations in genes other than RYR1. The mutational approach in our study has some limitations. First, we can not rule out deep intronic or non-canonical splice site variants, since the primers were designed to cover only exons and canonical splice sites. With our methods the non-canonical splicing events, that are emerging to be numerous and often tissue specific, have been missed and therefore also their potential role in core myopathies may be under recognized. Second, we cannot exclude to have missed some RYR1 variants because of some intrinsically pitfalls in Sanger sequencing (i.e. the quality of sequence in the first 15 to 40 bases where the primers bind, or quality of sequencing in long reads, etc.), and finally, RYR1 Sanger sequencing does not allow to look for mutations in other known genes possibly involved in core myopathies. The knowledge of all deleterious alleles in a given patients may contribute to the understanding of the role of the mutational load in disease phenotype expression.
The cohort of patients studied confirmed clinical heterogeneity in RYR1-related core myopathy, even if the selection criteria we used to enrol patients in this study were chosen to obtain a homogeneous group of patients adding to the presence of RYR1 mutations a histopathological diagnosis of core myopathy.
RYR1-related AD CCD presents with a variable clinical spectrum including congenital (47%), early (38%) and adult (14%) onset manifestations. A delay in reaching motor milestones was frequently described (30 patients), Page 15 of 20 Fusto et al. Acta Neuropathologica Communications (2022) 10:54 and only one patient is still not walking at the last evaluation at 2 years of age. In patients with adult onset, symptoms and signs at clinical presentation included cramps and myalgias, fatigue and muscle weakness prevalent in the lower limbs. Distribution of weakness was symmetrical, both proximal and distal, in the pelvic and shoulders girdles in 19% of patients and in a minority included also axial muscle weakness. Facial weakness, configuring the typical myopathic face in congenital myopathy, was present in 35% of patients. Spine and joint contractions were frequent in our cohort.
As previously reported [56,62,106], patients with recessive RYR1 mutations (8) were overall more severe than AD cases. Onset was congenital in all, arthrogryposis (37.5% vs 0%), floppy infant presentation (62.5% vs 17%), respiratory insufficiency (37.5% vs 0%) and hip dislocation (29% vs 15%) at onset were more frequent in AR than AD cases (Additional file 3: T2). However, even in the AR cases disease progression was variable (stable in 4 patients, slowly progressive in 3 and one patient showed improvement over time). At last evaluation (range 3-40 years) all patients were ambulant but one, who was able to walk only with support at 16 years of age. Overall disease phenotype was severe/moderate with frequently associated foot deformities (50%) and scoliosis (75% of patients).
Thirty-one patients in our cohort were isolated cases. Unfortunately, parents were not available to unequivocally establish the inheritance pattern. Based on the inheritance pattern of previously reported mutations 3 cases were likely AR CCD (i.e. patient #1 carrying two mutations previously reported in trans [54]; patient #9 [62] and 38 [82] carrying a RYR1 mutation previously reported in AR cases in compound heterozygosity [62,82]), and 6 likely AD CCD (i.e. patient #8 [61], #14 carrying a mutation previously reported in severe AD centronuclear myopathy [70], #29 [77], #30 and #33 harbouring mutations previously reported in core/rod myopathy [75] and in congenital myopathy with uniformity of type 1 muscle fibres [80] and patient #42 [83]). Three patients carried RYR1 mutations previously associated to AD MH (i.e. #2 [55], #16 [57] and #26 [55]) (Table 1; Additional file 3: T2). In the remaining cases either novel mutations or previously unreported association of various RYR1 mutation do not allow to infer the phase of the variants. All 3 putative AR cases had a congenital onset and mild to severe myopathy associated with contractures or scoliosis according to the rule of more severe phenotype in AR vs AD cases.
It is interesting to note that at least 15 patients in our cohort carried mutations associated to MH. MH is usually linked to minimal or absent muscle weakness and muscle biopsy may, or may not, show cores. Given our selection criteria, all patients presented cores at muscle biopsy and 3/15 showed moderate and 6/15 mild muscle weakness. Two patients (#16 and #26) presented a congenital onset (with arthrogryposis, floppiness, hip dislocations in patient #16 and foetal hypokinesia, dysphagia in patient #26), delayed independent ambulation (30 and 24 months respectively) and an overall mild myopathy with stationary progression over time but associated with joint contractures, scoliosis and pes cavus. In these two patients, given the severe clinical picture, it is likely that a second RYR1 mutation was not detected. Only two patients were asymptomatic (in one diagnosis was reached because of a positive family history and in the second in the work up of a hyperCKemia) and two were paucisymptomatic (myalgia, cramps). Among patients carrying mutations previously associated to MH, at least three had a MH reaction during anaesthesia: patient #3 during pes cavus surgery, patient #21 during surgery for congenital muscular torticollis and patient # 23 during spinal arthrodesis for scoliosis. Patient #3 and #21 had an early onset moderate myopathy with typical features of congenital myopathy, where patient #23 was paucisymptomatic and was studied for hyperCKemia. Given the selection bias in our cohort of patients we were not able to assess either the percentage of MH patients presenting with muscle weakness nor the percentage showing core at histopathology. It is possible to speculate that the presence of core at muscle biopsy predisposes to the development of muscle weakness, but a large number of patients is needed to reach a definite conclusion. On the other hand, patients in our cohort may simply express the clinical variability associated to RYR1 mutations. For example, family 3, carrying the well-known p.Arg163Cys mutation associated with MH, displayed a wide phenotypic spectrum. Among the members of this family, one individual (patient #3) presented with an early onset moderate myopathy with muscle wasting, contractures, foot deformities and scoliosis, patient#4 had an adult onset mild myopathy, while patient #5 came to medical attention for asymptomatic hyperCKemia.
The reasons of this intrafamilial variability are not known but genetic modifiers in genes other than RYR1 may be responsible for the clinical variability.
Five patients (belonging to four families) in our cohort carried heterozygous small frame-shift deletion or insertion mutations causing a premature stop codon predicted to result in a reduced level of RyR1. Recently, it has been shown in a knocked-in mouse for a RYR1 single allele frame-shift mutation, that the level of RYR1 transcript and RyR1 protein are decreased without affecting the other SR proteins. The mouse also showed a mild reduction of muscle performance and decreased muscle . Indeed, all six patients followed the postulated rule of a mild reduction in muscle strength.
The RYR1 gene carries a large number of polymorphisms as well as causative mutations: there are 1,783 public variants vs 843 unique public variants reported in LOVD database (https:// datab ases. lovd. nl/ shared/ genes/ RYR1; accessed June 2021) making the assessment of pathogenicity of the identified variant a crucial issue. We identified 68 RYR1 variants and 16 were novel.
We used an integrative in silico approach to evaluate the pathogenic impact of some of the RYR1 mutations identified. Our approach included an in-depth literature search along with a systematic bioinformatics analysis of some of the RYR1 variants identified that allowed us to determine the molecular details supporting the pathogenicity of several novel RYR1 mutations. Despite our promising data, several hurdles have to be considered: the lack of the human RyR1 crystal structure paired with the presence of large intrinsically disordered regions, the large size of the protein and its composition of multiple functional and regulative elements, all pointing to a difficult assessment of a single variant.
In heterozygous patients the protein may be constitute of variable proportion of mutated monomers, the ratio wild-type/mutated monomers could be directly linked to the disease severity but assessing stoichiometry in single patient it is not feasible at the moment. On the other hand, different mutations lead to different pathogenic mechanisms, such as haploinsufficiency, inactivation of regulative sites, impairment of complex assembly and alterations in the channel's gating properties. To add to the complexity, the genetic background may modify the pathological effect of the RYR1 mutations. For example, variations in other genes associated with core myopathies or known to be RyR1 interactors (i.e. FKBP1B, TRDN, ASPH, FKBP1A, STAC3, CACNA1S, CACNA1C, CACNA1A, NOS1, CALM1) or even other unknow genes may present single nucleotide polymorphism that cause subtle modification on the protein product but sufficient to modify the phenotype. This hypothesize phenomenon is well represented by the phenotypical variability among member of the same family. Finally, as mentioned before, epigenetic modifications of RYR1 may also play a role [108] or the methylation status of genes implicated in cytosolic Ca 2+ buffering or trafficking of the Na + /Ca 2+ exchanger [109].
The difficult interpretation of the pathogenicity of RYR1 variant is reported also by Johnston and colleagues, who focused their attention on RYR1 mutations causing malignant hyperthermia. They revised the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMP/AMP) criteria for variant interpretation highlighting the difficulty to define general rules to estimate the effect of a mutations [110].
From a myopathological point of view the majority of patients (67%) showed at muscle biopsies core lesions in type 1 muscle fibres, 17% of patients had multiple minicores in both type 1 and type 2 muscle fibres and 15% of muscle biopsies showed both central core and multiple minicores. In a subset of patients, the presence of cores and minicores was confirmed with ultrastructural electron microscopy (EM) studies but EM was not formally investigated in all patients. No correlations were detected among the number of fibres with core and clinical severity. It is interesting to note that the only patient who had 2 muscle biopsy done at age 5 and 12 years showed a dramatic modification of muscle morphology. The first muscle biopsy did not show cores while the biopsy taken at 12 years of age disclosed central core in 100% of type 1 fibres. These observations are in line with the hypothesis that core lesions may not represent a primary developmental abnormality but could be secondary to a maintained abnormal contraction caused by any excitation-contraction coupling defect as previously reported [23].

Conclusions
Considering all the limitations detailed above we were not able to correlate specific genotype to phenotype, but focusing on protein domains we suggest that, in our cohort, mutations in some domains are more frequently associated with specific phenotypes. For example, the pore domain is associated with foetal hypokinesia, contractures and foot deformities. Moreover, it is of interest that all patients (#48 and #65) carrying mutations in Ry1&2 presented with particularly severe phenotype, as this domain is associated with foetal hypokinesia, bulbar involvement and respiratory insufficiency. However, the low number of patients do not allow definitive conclusions at this point.