Central Core Disease: Facial Weakness Differentiating Biallelic from Monoallelic Forms

Central Core Disease (CCD) is a genetic neuromuscular disorder characterized by the presence of cores in muscle biopsy. The inheritance has been described as predominantly autosomal dominant (AD), and the disease may present as severe neonatal or mild adult forms. Here we report clinical and molecular data on a large cohort of Brazilian CCD patients, including a retrospective clinical analysis and molecular screening for RYR1 variants using Next-Generation Sequencing (NGS). We analyzed 27 patients from 19 unrelated families: four families (11 patients) with autosomal dominant inheritance (AD), two families (3 patients) with autosomal recessive (AR), and 13 sporadic cases. Biallelic RYR1 variants were found in six families (two AR and four sporadic cases) of the 14 molecularly analyzed families (~43%), suggesting a higher frequency of AR inheritance than expected. None of these cases presented a severe phenotype. Facial weakness was more common in biallelic than in monoallelic patients (p = 0.0043) and might be a marker for AR forms. NGS is highly effective for the identification of RYR1 variants in CCD patients, allowing the discovery of a higher proportion of AR cases with biallelic mutations. These data have important implications for the genetic counseling of the families.


Introduction
Central Core Disease (CCD) is one of the most common congenital myopathies. Patients' phenotypes range from a rarer severe neonatal myopathy similar to a congenital muscular dystrophy to a most frequent proximal weakness, manifesting in adulthood.
CCD was first described in 1956 as a congenital non-progressive myopathy characterized by hypotonia, delayed motor milestones, and a mild non-progressive weakness that predominantly affects proximal lower limb muscles [1]. This description is still very accurate, and reflects the clinical presentation of most dominantly inherited CCD patients [2].
Contrary to other congenital myopathies, CCD usually presents only discrete facial and bulbar weakness but prominent orthopedic complications such as congenital hip dislocation, Achilles tendon contractures, and joint hypermobility [3].
The histopathological hallmark of CCD is the presence of cores in the muscle fibers of the patients. Cores are areas with reduced oxidative activity observed through NADH, SDH, and COX reactions. They correspond to areas devoid of mitochondrial activity that appear on electron microscopy as round intrasarcoplasmic areas of myofibrillar disorganization with scarce or no mitochondria.
Most CCD patients present RYR1 gene pathogenic variants. The RYR1 gene is a huge gene with 106 exons, located at 19q13.2, with 5037 amino acids forming a 535 kDa polypeptide. This gene is also involved with the pathogenesis of other neuromuscular disorders such as Malignant Hyperthermia Susceptibility (MHS), Multiminicore disease (MmD), and Centronuclear myopathy (CNM). Malignant hyperthermia is a life-threatening condition triggered using volatile anesthetics such as halothane, and myorelaxants such as succinylcholine. Usually, CCD patients present both CCD and MHS.
The inheritance pattern of CCD is predominantly autosomal dominant, but both recessive and sporadic cases have been described [4]. The very rare severe neonatal form is characterized by the presence of profound hypotonia and has been related to RYR1 hypomorphic variants that produce reduced levels of total RyR1 protein. These cases have been associated with recessive missense/indel RYR1 variants [5].
The identification of RYR1 gene variants is very challenging as more than 450 variants have been identified causing both CCD and MHS. These variants are distributed mainly in three hotspots: sarcoplasmic D1 that includes N-terminal residues 1-614, sarcoplasmic D2 that includes central region residues 2163-2458, pore-forming, SR lumen, and membrane D3 that includes the C-terminal residues 4136-4973. There is a difference between MHS and CCD predominant hotspot regions: MHS in D1 and D2, and CCD in D3 [5]. Nextgeneration sequencing (NGS) is a revolutionary method for RYR1 gene study. Before NGS, the screening of RYR1 variants was performed by Sanger Sequencing, and predominantly in the hotspot regions due to the large size of the gene and the broad distribution of the pathogenic variants. NGS has proved to be a more efficient and cost-effective technique to screen for mutations because, in addition to the RYR1 gene, many other genes and regions can be analyzed at the same time. NGS panels may be customized according to the studied population.
The objective of this study is to describe the clinical variability in a series of Central Core Disease patients from one reference center and to correlate them with the RYR1 gene variants and the pattern of segregation. The results have important implications for genetic counseling and for the study of physiopathological mechanisms involved in CCD.

Materials and Methods
A retrospective case series study of medical records from 1997 to 2019 has been performed, including 27 patients with Central Core Disease belonging to 19 unrelated families. Some of these patients were previously included in publications focusing on different aspects of CCD. Here, patients' inclusion criteria were based on data from muscle biopsy with proven diagnosis of Central Core Disease and their symptomatic relatives.

Patients Clinical Evaluation
Medical charts review was performed with the evaluation of relevant clinical, laboratory, imaging, and muscle biopsy data.
Clinical characteristics included: age at first symptoms; age at diagnosis; gender; inheritance pattern; consanguinity; oligohydramnios; fetal akinesia; delayed motor milestones; hypotonia; bulbar weakness (sucking/swallowing); cardiac and respiratory evaluation, neonatal severe respiratory involvement; congenital hip dislocation; and club feet history.
Ancillary studies included electromyogram, serum total creatine kinase and aldolase levels, echocardiogram, muscle imaging, and muscle biopsy. Muscle imaging included the study of pelvis and lower limb muscles on T1-weighted axial sequences or muscle Computed Tomography in cases in which MRI could not be performed. Muscle biopsy analysis included the study of liquid nitrogen frozen specimens analyzed through the mitochondrial oxidative reactions, Succinate dehydrogenase, Cytochrome-c-oxidase, Nicotinamide adenine dinucleotide (NADH), and other routine techniques: hematoxylin and eosin, modified Gomori trichrome, Periodic acid Schiff (with and without diastase), Oil-red-O, myosinic ATPase (pH 9.4, pH 4.6, and pH 4.3), acid phosphatase, and nonspecific esterase.

Molecular Analysis
Patients' DNA samples were extracted from peripheral blood lymphocytes using a routine methodology.
The genetic investigation was performed firstly by using an NGS customized panel for 95 neuromuscular diseases (NMD) genes including the RYR1 gene. Afterwards, we used the llumina TruSight One Expanded panel, which targets more than 6700 genes and exonic regions associated with clinical phenotypes.
The following preparation kits used were: SureSelect OXT library and SureSelect Human all exons and V6 capture kit (Agilent, Santa Clara, CA, USA). The sequencing was performed on Hiseq2500 equipment (Illumina, San Diego, CA, USA). The human genome reference for data alignment was the version GRCh37/hg19.
A control population of 1000 Genomes, NIH, gnomAD, 6500 Exaomes Sequencing Project (Washington University, Washington, DC, USA), and the new AbraOM (Online Archive of Brazilian Mutations) were used for comparison with the filtered variants. The RYR1 gene (OMIM#180901, transcript NM_000540.3) rare variants were checked and analyzed using bioinformatic tools. Gene Mutations Databases HGMD, LOVD, and ClinVar were checked for already described pathogenic RYR1 variants. The American College of Medical Genetics and Genomics (ACMG) pathogenicity classification guidelines [6] were used for variants classifications. The prediction of pathogenesis of de novo variants was supported evaluating the results from many in silico prediction softwares: MutationTaster, Predict SNP1, CADD, DANN, FATHMM, FunSeq2, GWAVA, VEP, SIFT, Polyphen2, and Human splicing finder 3.0.
For the confirmation of pathogenic variants, relatives screening, variants segregation study within the families, and Sanger Sequencing of specific exons were performed.

Data Analysis
Data analysis was performed with approval of the institutional Scientific and Ethical Committees. Statistical descriptive analysis and comparison between clinical groups were performed using the Fisher exact test with statistical significance for p < 0.05.

Results
A total of 27 patients from 19 families were selected. Sporadic cases were the most common, being half of the cases (13/27 patients), followed by autosomal dominant (11 patients from four families) and recessive inheritances (three patients from two families) ( Figure 1A). Imaging studies were used to guide the choice for the best site for muscle biopsy. Due to the frequent muscle fat replacement of the vastus lateralis, the rectus femoris was the muscle of choice whenever the former was involved ( Figure 2A).
All patients submitted to muscle biopsy presented areas devoid of oxidative reaction correspondent to myofibrillar disorganization with scarce mitochondria ( Figure 2B,C). Imaging studies were used to guide the choice for the best site for muscle biopsy. Due to the frequent muscle fat replacement of the vastus lateralis, the rectus femoris was the muscle of choice whenever the former was involved (Figure 2A). Imaging studies were used to guide the choice for the best site for muscle biopsy Due to the frequent muscle fat replacement of the vastus lateralis, the rectus femoris was th muscle of choice whenever the former was involved ( Figure 2A).
All patients submitted to muscle biopsy presented areas devoid of oxidative reactio correspondent to myofibrillar disorganization with scarce mitochondria ( Figure 2B,C). A summary of clinical, laboratory, and molecular data is provided in Tables 1-3. Six teen patients were female and eleven were male. Most patient presented hypotonia (14/23 and developmental delay (15/23) ( Table 1). All patients submitted to muscle biopsy presented areas devoid of oxidative reaction correspondent to myofibrillar disorganization with scarce mitochondria ( Figure 2B,C).
A summary of clinical, laboratory, and molecular data is provided in Tables 1-3. Sixteen patients were female and eleven were male. Most patient presented hypotonia (14/23) and developmental delay (15/23) ( Table 1).
The mean age at diagnosis was 18 ± 15.8 years and the age range varied from 1 to 52 years. Only one patient reported first symptoms in adulthood (Table 1). All other patients presented first symptoms in childhood.
Congenital hip dislocation was observed in 5/25 patients and congenital club feet in 4/26 patients. Bulbar symptoms, characterized by difficult sucking or swallowing, were observed in 3/15 patients and severe neonatal respiratory involvement was reported in 2/17 patients.
No signs of significant cardiac or pulmonary abnormalities were observed: all patients presented normal cardiac and pulmonary auscultation. Nevertheless, all of them received recommendation to be submitted to cardiac evaluation at external specialized services.  Facial weakness was observed in 8/27 patients and facial dysmorphisms, characterized by long face, high arched palate, and dolichocephaly were observed in 6/27 patients. Palpebral ptosis was observed in 2/28 patients and none of them presented ophthalmoplegia.
Serum total creatine kinase levels were normal or almost normal in 84% (22/26) of the patients. One patient (18.1 in Table 2) presented 4.3 times (966 IU/L) increased total creatine kinase levels ( Table 2).
A total of 21 variants in the RYR1 gene were identified in the 14 molecularly studied families (five sporadic cases were not available for the molecular analysis). Two variants were recurrent: p.Ala4846Val in two families (#5, #6) and p.Arg4861His in three families (#8, #10, #13) ( Table 3).
Right calf atrophy, winged scapula Gait with orthesis at age 5  Among the 18 different mutations, 16 were missense nonsynonymous variants, one was a nonsense variant, and one was an intronic variant. Fourteen of the variants were previously described as pathogenic [5,[7][8][9][10][11][12][13][14][15][16][17][18] and four variants are now being associated with the CCD phenotype. Eleven of the mutations were localized in the C-region, while three were in hotspot D1 and D2 N-terminal and sarcoplasmic domains of the RyR1 channel, and four variants were outside of any of the three main domains (Figure 3). Genes 2022, 13, x FOR PEER REVIEW 8 of 13 Molecular analysis disclosed monoallelic variants in the four AD families and in four sporadic cases, while biallelic variants were found in the two AR families (three patients) and in four sporadic cases ( Figure 1B). Therefore, genotypic analysis disclosed 8/14 (57%) families with monoallelic variants (15 patients: 11 AD and four sporadic) and 6/14 (43%) families with biallelic variants (seven patients: three AR and four sporadic). Only one family was consanguineous; however, all the biallelic mutations were compound heterozygous (Table 3).
A comparison between the phenotypes found in monoallelic versus biallelic patients was performed. There were no statistically significant differences between both groups considering the frequency of facial dysmorphism, axial weakness, distal weakness, hypotrophy, bulbar symptoms, congenital club feet, severe neonatal respiratory distress, or hip dislocation (Table 4). However, facial weakness was significantly more frequent in the biallelic (71.0% or 5/7) than the monoallelic (6.7% or 1/15) patients (p = 0.0043) (Figure 4) ( Table 4).  Molecular analysis disclosed monoallelic variants in the four AD families and in four sporadic cases, while biallelic variants were found in the two AR families (three patients) and in four sporadic cases ( Figure 1B). Therefore, genotypic analysis disclosed 8/14 (57%) families with monoallelic variants (15 patients: 11 AD and four sporadic) and 6/14 (43%) families with biallelic variants (seven patients: three AR and four sporadic). Only one family was consanguineous; however, all the biallelic mutations were compound heterozygous (Table 3).
No patient presented any signs of respiratory involvement as a characteristic clinical finding of Central Core Disease [26].
Pathogenic variants in the ryanodine receptor (RYR1) gene are the most common causes of congenital myopathy, and the most frequent cause of Central Core Disease [21,27]. With the introduction of NGS testing and the possibility to screen all the 108 exons of the RYR1 gene, the capacity to detect mutations was significantly improved. In fact, although a predominance of variants in the C-terminal D3 domain (11/18 variants) is still observed, three variants were localized in domains D1 and D2 and an additional four variants (two novel) were found outside of these three hotspot domains. These findings strengthen the knowledge that CCD mutations are indeed located in the D3 domains, but also point to the presence of pathogenic variants outside these regions, showing the importance of sequencing the whole RYR1 gene to identify such variants that would have been missed if only the hotspots regions were sequenced.
Genetically, most of our CCD patients present autosomal dominant mild, non-progressive limb weakness, as already described [21]. Nevertheless, although severe CCD with respiratory and bulbar symptoms was usually described in recessive cases [21], in this series, only 3/18 patients presented severe bulbar symptoms, and their inheritance pattern was either autosomal dominant or sporadic. Two severe cases (16.1 and 19.1) were not submitted to molecular analysis.
Additionally, none of the patients from our six families with biallelic variants presented a severe phenotype with bulbar weakness. These results suggest that with the augmentation of the capacity of the molecular identification of CCD cases with biallelic mutations, the clinical phenotype is exhibiting a broader variability, with milder cases described recently.
As to the genotype-phenotype correlations, there were some clinical and morphological differences between the patients here described and previous publications of the same RYR1 variants. The p.Arg401His and p.Arg614Cys variants were previously described as monoallelic, causing disease mutations in patients with malignant hyperthermia [16,17]. We present here a case of a patient (12.1) in which the combination of these two variants causes a Central Core Disease phenotype, being an example of phenotypic variability associated with RYR1 mutations.
The p.Ala4846Val (exon 101) variant was previously described both in Central Core Disease and in an autosomal recessive centronuclear RYR1-related congenital myopathy [7,11]. Both conditions presented with developmental delay and hypotonia. However, facial weakness without ophthalmoplegia was described in Central Core Disease [7], while ophthalmoplegia without facial weakness was reported in RYR1-related centronuclear myopathy [11]. The variant p.Gly4897Asp was previously described both as a monoallelic and biallelic disease [5,28], meaning that this variant can cause a dominant phenotype when present in heterozygosis, probably associated with a second genetic modification that predisposes to the disease, but also a recessive phenotype when in homozygosis. Therefore, clinical variability can be detected in patients carrying the same mutations, suggesting the action of other factors modulating the phenotype.
Regarding the pathogenicity of the novel mutations here identified, three out of four were classified as likely pathogenic (if we assume that they are de novo variants but without confirmation of paternal and maternal segregation): the variant p.Asn4834Ty (PM2_Strong, PM6, PP3), the variant p.Asp4816Asn (PM2_Strong, PM5, PM6, PP3), and the variant p.Ala3431Val (PM2_Strong, PM6, PP3). The variant p.Ala3431Val was found only in Patient 13.1, who also has the pathogenic p.Arg4861His described mutation found in heterozygosis in another two of our sporadic patients (8.1 and 10.1), as well as in another report [13]. This suggests that, although the novel variant p.Ala3431Val might be pathogenic, the main mutation causing CCD in the patient is p.Arg4861His. The remaining fourth novel mutation p.Gln1613Ter causes a truncated protein so, as expected, it received a pathogenic classification (PVS1, PM2, PP3, PP5). This variant was found in association with a second previously described mutation p.Glu1175Lys [15], which was classified as a VUS (Variant of Uncertain Significance) by the authors. Once the p.Gln1613Ter mutation is sporadic, we cannot conclude whether it alone is sufficient to cause a CCD phenotype in a dominant manner, or if the presence of another mutation is mandatory.
Interestingly, Patient 6.1 has the same two mutations (p.Gly2343Ser and p.Arg4558Trp) that were identified in three different patients [10,11] also in compound heterozygosis with an autosomal recessive inheritance. As with our patient, both patients reported by Samões, 2017 also had a third mutation at RYR1 [10].
A comprehensive multiprofessional orthopedic and genetic evaluation of these patients may provide valuable clinical orientation and genetic counseling..
Comparing mono versus biallelic patients, only facial weakness was able to differentiate both groups with statistical significance. This important clinical signal can thus direct the analysis of new cases, aiming at genetic counseling.
Finally, in this large cohort of patients from the same reference center, it was not possible to infer that biallelic cases presented more severe phenotypes than monoallelic cases. On the other hand, it provided some evidence that facial weakness might be a clue to biallelic inheritance.

Conclusions
NGS is improving our capacity to identify mutations in the RYR1 gene and increasing the number of identified pathogenic mutations, either as a heterozygous allele or compound heterozygous alleles. This is the probable reason for the increasing number of biallelic cases than previously reported. Facial weakness was more common in biallelic than in monoallelic patients in this group, and it could be a marker for AR forms. These data have important implications for the genetic counseling of the families.