The research protocol (2015[916]) was reviewed and approved by the Ethics Committee of Peking University First Hospital. Written informed consent, which also included consent for the publication of medical data, was obtained from the patients and their parents.

Medical records of the patients were reviewed. Clinical and laboratory data, including age of onset, onset of recognizable symptoms, motor development, contractures, creatine kinase (CK) levels, electrocardiography (ECG), Holter, ultrasound cardiography (UCG), cardiac magnetic resonance imaging (MRI), limb muscle MRI, and muscle biopsy findings, were collected.

Genomic DNA of the proband, the younger brother, and their parents was extracted from leukocytes isolated from peripheral blood. A muscular disease gene panel and whole exome sequencing (WES) were successively used to detect genetic variations in the proband. Candidate variants in the probands were validated by Sanger sequencing in the family. Disease association databases and population databases, including the HGMD (HGMD: http://www.hgmd.cf.ac.uk/ac/index.php), LOVD (LOVD: http://www.dmd.nl/), ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/), 1000 Genomes (http://www.1000genomes.org/), ExAC (ExAC: http://exac.broadinstitute.org/), and gnomAD (gnomAD: http://gnomad.broadinstitute.org/) databases, were used to identify previously reported mutations and discover potential novel mutations. Predictions for the pathogenicity of missense mutations were obtained using PolyPhen-2, SIFT, and Mutation Taster. The predicted splicing mutations were tested through the Human Splicing Finder. The pathogenicity of candidate variants was classified according to the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG-AMP) guidelines (Richards et al., 2015).

Next-generation sequencing (NGS) data were also analyzed to detect copy number variations (CNVs) by comparing the number of sequence reads between the patient and control samples. First, the number of reads in each amplicon was counted for the patient and control samples. For the purposes of this comparison, sequencing data from at least three control samples (genomic sequencing had been performed for patients diagnosed with other diseases) were used as the reference. Subsequently, the patient/control ratios were calculated by dividing the number of patient reads in each amplicon with the average of the total reads regarding this amplicon from control samples. For the genes, including EMD and its neighboring gene FLNA, located on the X chromosome, patient/control ratios near 0.5 indicated the presence of heterozygous deletions in female patients (control samples were female). Ratios near 0 indicated deletions in male patients (control samples were male). In contrast, ratios near 1.5 indicated the presence of heterozygous duplications in female patients (control samples were female), and ratios higher than 1.75 indicated duplications in male patients (control samples were male). Polymerase chain reaction (PCR) was used to confirm the deletion of EMD in male patients. Statistical results were illustrated using GraphPad Prism 8 software (GraphPad Software, La Jolla, CA).

PCR and Sanger sequencing were performed to verify EMD deletion and to sequence the breakpoints of the CNVs. Primers were designed separately for exons 1–6 of EMD using Primer 5.0 (Premier Biosoft, Palo Alto, CA, United States). The two primers designed for exon 1 were Primer F 5′-ACC​GCG​AGA​CCT​TTT​GCT​C-3′ and Primer R 3′-GCG​AGG​CTC​TCA​CCA​GAG​AA-5′. The two primers designed for exon 2 were Primer F 5′-CGC​ACG​GGC​CTG​TAG​TAG​GT-3′ and Primer R 3′-GAC​GAA​GTC​GGG​TGA​AGG​TG-5′. The two primers designed for exon 3 were Primer F 5′-CGG​CCA​GGA​TCA​ACT​CGT​AG-3′ and Primer R 3′-TCC​TAA​GGC​TGC​TGG​AGT​GG-5′. The two primers designed for exon 4 were Primer F 5′-ACC​TTC​ACC​CGA​CTT​CGT​CA-3′ and Primer R 3′-TCC​TGA​ATG​GCT​CTG​GGT​GT-5′. The two primers designed for exon 5 were Primer F 5′-GAG​CCA​TTC​AGG​AGG​GTG​TG-3′ and Primer R 3′-GCT​GAA​ACA​GGG​CGG​TAG​TG-5′. The two primers designed for exon 6 were Primer F 5′-TGT​GGG​TTC​CTG​GCC​TCT​AA-3′ and Primer R 3′- AAA​AAT​GCG​GAC​CCA​ACA​AA-5′. For the breakpoints, two primers were designed for a location in exon 2 of FLNA (Primer F GRCh37/hg19, chrx:153599438-1535994, 5′-CAC​TTC​AGG​TGC​TCG​TT-3′) and exon 29 of FLNA (Primer R GRCh37/hg19, chrx:153585865-153585883, 3′-GGA​TCT​CGT​CAC​CAC​CGT​A-5′). PCR amplifications were performed using GoldStar Taq DNA Polymerase in an Applied Biosystems Veriti Thermal Cycler (4375305, Thermo Fisher Scientific, Waltham, MA, United States). Each PCR amplification was repeated for 35 cycles (30 s at 95°C, 30 s at 60°C, and 45 s at 72°C). Before sequencing, the PCR products were analyzed by electrophoresis on a 2% agarose gel.

The proband was referred to our hospital at the age of 17 due to suspected “skeletal muscle cardiomyopathy”. He started to feel physical activity-related tiredness and chest distress at the age of 13. He was suspected of having “myocarditis” and was given nutritional therapy without regular follow up. At the age of 15, he visited the local hospital again due to poor exercise tolerance and chest discomfort. Subsequently, bradycardia and enlargement of the right heart were found by Holter monitoring and UCG. Due to the symptomatic right heart problem, he was suspected of having “right ventricular cardiomyopathy” and received indolapril as a treatment. At the age of 16, he presented with bilateral brachial muscle pain and aggravated chest distress. His motor and cognitive development was normal. None of his family members exhibited any clinical abnormalies. Physical examination at the age of 17 revealed a heart rate of 56 bpm, normal oxygen saturation, and no dyspnea at rest. Neurological examination showed obvious bilateral brachial and peroneal muscle atrophy, elbow contracture, a mild rigid spine, and a tight Achilles tendon with normal gait. The muscle strength grade of the upper proximal arms was IV⁺, while that of the distal arms was V. The muscle strength grade of the proximal lower limbs was V^−, with a distal dorsal flexion power of IV and a plantar flexion power of V⁻. Bilateral knee tendon reflex could not be induced. To confirm the diagnosis, laboratory tests were performed. The level of CK was 500–2,500 U/L (normal 0–200 IU/L). Skeletal muscle biopsy indicated muscular dystrophic changes (Figure 1A). Limb muscle MRI showed infiltration of fatty tissues, predominantly in the soleus (Figure 1B). UCG showed enlargement of the right heart with a left ventricular ejection fraction (LVEF) of 68% (Figure 1C), and the Holter monitoring showed a junctional ectopic rhythm of 36 bpm (Figure 1E). A cardiac MRI subsequently revealed a markedly enlarged right heart, evident thinning of the lateral wall of the right ventricle, and fatty tissues in the local cardiac muscle with LVEF of 57% (Figure 1D). Combining the predominant cardiac involvement, muscle weakness, and joint contractures, he was clinically diagnosed with EDMD. This was confirmed by genetic tests (listed below). Considering bradycardia and the risk of sudden cardiac death, he was referred to implant a cardiac pacemaker. In addition, he was suggested to undergo rehabilitation training with regular follow up.

A muscular disease gene panel that was performed in a local hospital detected no pathogenic variation in the proband. We reanalyzed the NGS data. We did not find any pathogenic variation but found that the depth of coverage was low in the whole region of EMD. This suggests EMD deletion. The PCR results confirmed the entire deletion of EMD (exons 1–6) (Figure 2A) in the proband. Then, we needed to further detect the deleted region and breakpoints. Due to the lack of designed probes and the small size of EMD, we failed to analyze the deleted region by using multiplex ligation-dependent probe amplification (MLPA) and array CGH. Finally, we used the NGS data of WES combined with PCR to help verify the region of deletion. The depth of coverage was used to detect the CNV by comparing its copy number ratio between the patient and the control. The ratio was 0 in the region of EMD, suggesting the deletion of the entire EMD gene (Figure 2B). On the other hand, the ratios were almost all greater than 1.75 in one segment (exon 29_48) and approximately 1 in another segment (exon 2_28) of FLNA, a gene adjacent to EMD. These findings suggest partial duplication of FLNA corresponding to exons 29–48 (Hg19, NM_001110556, Figure 2B). However, due to the lack of reads designed to cover exon 1 of FLNA, the WES data were lost in FLNA exon 1, which is a non-coding region (Patrosso et al., 1994). The whole EMD deletion involving partial FLNA duplication was similar to the genomic arrangement of the first reported EDMD case with EMD deletion (Small et al., 1997). Therefore, according to the case reported before (Small et al., 1997), we proposed that the genomic rearrangement in the proband may be caused by nonhomologous end joining on the background of FLNA-EMD inversion. Considering that the duplicated region of FLNA (exon 29_48) starts from exon 29, we proposed that one breakpoint in FLNA was located in intron 28 of FLNA. Besides, FLNA exon 2 was unaffected. Therefore, another breakpoint in FLNA may occur in intron 1 of FLNA to the linkage region of FLNA-EMD. A model including the inversion of FLNA-EMD is shown (Figure 3) to explain the rearrangement of FLNA. To verify the breakpoints, PCR and Sanger sequencing were performed. Two primer sets were designed to be located in exon 2 of FLNA and exon 29 of FLNA. After sequencing, deletion boundaries and sequences were found. Between homologous X or sister chromatins, a microhomology-mediated nonhomologous end joining event was confirmed, accompanied by a 5 bp overlap (GTCCC) between FLNA intron 28 and FLNA intron 1 on the background of FLNA-EMD inversion.

The family pedigree is shown in Figure 2C. The family members were then evaluated. The 14-year-old younger brother complained of a mild bowing limitation. His CK level was 549 U/L. His echocardiograph and electrocardiograph revealed no problems. Their parents showed no symptoms, with normal CK and cardiac test results. After the same genetic test, the younger brother was identified as a patient with EMD deletion and partial FLNA duplication. He was followed up and given rehabilitation training regularly. The mother was confirmed as a carrier.