Diagnostic Challenges of Neuromuscular Disorders after Whole Exome Sequencing

Patient recruitment

This retrospective cohort study (202109079RIBN) was approved by the Ethics Committee of the National Taiwan University Hospital (NTUH), Taipei, Taiwan. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The inclusion criteria were: (1) an impression of hereditary NMDs, (2) no other etiology identified after selective genetic testing before the WES study, and (3) clinical data were available. We arranged WES to all the probands, and for the proband’s family, Sanger sequencing on the specific variants revealed by the proband-only WES was arranged if they agreed.

Clinical evaluations, laboratory study, and electrophysiological examinations

The clinical data, including age of disease onset, sex, and pedigree, were collected in all patients. In patients with myopathy, we additionally collected the serum creatinine kinase (CK) levels and the informations of muscle histopathology. According to previous studies, distribution of weakness and associated symptoms can be categorized into different patterns for clinical diagnosis. The presentation of our patients was categorized as follows: (1) limb-girdle weakness, (2) scapuloperoneal weakness, (3) distal arm and proximal leg weakness, (4) distal weakness, (5) ptosis with or without ophthalmoplegia, (6) episodic weakness/myalgia/rhabdomyolysis, and (7) asymptomatic hyperCKemia. The muscle histopathology was evaluated by the neurologist and pathologist together. Routine histochemical staining in our hospital encompassed hematoxylin and eosin staining, Gomori-Trichome staining, succinate dehydrogenase staining, nicotinamide adenine dinucleotide staining, and staining for myosin adenosine triphosphatase at pH 4.3 and 10.42. Based on clinical information and routine histochemistry results, specific immunohistochemistry would be arranged. Electron microscopy studies of muscle biopsy were also performed on some samples.

In patients with polyneuropathy, we collected clinical data and performed nerve conduction study (NCS). We used the electrophysiological criteria for chronic inflammatory demyelinating polyradiculoneuropathy to define demyelination and axonal loss.

Whole-exome sequencing

WES was performed using an exome capture kit probe (Agilent V6 (before Jan 2021) and Roche Hyperexome (after Feb 2021)). We then did sequencing with a NovaSeq6000 sequencer (Illumina). The protein-coding genes and flanking intronic sequences (50 bp) were studied with a 300-bp paired-end run and an average of >  150-fold coverage.

The human reference genome is GRCH38 and we used the Burrows–Wheeler Aligner (BWA) for sequence alignment and used the Genome Analysis Tool Kit (GATK V3.5, Broad Institute) [13] for variant calling. Variants were first annotated using ANNOVAR (http://wannovar.wglab.org/) [14]. Then, the annotated profiles were tagged with information of the inheritance pattern from OMIM (https://omim.org/), variant pathogenicity from ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), and the allele frequency from the Taiwan Biobank (https://taiwanview.twbiobank.org.tw/). The variant filtering criteria incorporated maximal minor allele frequency (included if <  0.01) and in-silico analysis of variant severity (nonsynonymous variant, nonsense, deletion, splicing, insertion, or splicing). We manually calssified the pathogenicity of variants according to American College of Medical Genetics and Genomics (ACMG) and Association for Molecular Pathology guidelines [15–19]. We further defined the results of WES as “positive”, “inconclusive” and “negative” based on ACMG criteria. The positive results were defined as the variants that had been asserted as pathogenic or likely pathogenic in ACMG criteria and definite genetic diagnosis according to the inheritance pattern of gene. That is, for those compound heterozygous variants, only both alleles showed pathogenic or likely pathogenic in ACMG criteria were defined as “positive”. The inconclusive results were those being classified as variants of uncertain significance in ACMG criteria. The results of the WES were defined as negative if (1) no possible culprit variant identified in the WES or (2) the identified variants were excluded by the segregation analysis. For copy number analysis (CNV), with SAM/BAM file as input, the GATK germline CNV (gCNV) caller can generate a VCF file containing the CNVs found during the variant calling process [20].

Clinical validation of WES results: Clinical-Molecular diagnosis

In patients with myopathy, the pathogenicity was defined as “definite” in two situations: (1) patients presented with the representative clinical phenotype with positive WES results, or (2) those who had a typical phenotype but with inconclusive WES results, the diagnosis validated by typical pathognomonic findings in clinical manifestations, pathology, or other examinations. Notably, there should be negative results for the other known genes with the similar phenotype in both situations [21]. Instead, pathogenicity was defined as “probable” if the criteria were not met. Among the patients with neuropathy or motor neuron disease, the clinical-molecular diagnosis was regarded “definite” only if the clinical manifestations, culprit variants, and electrophysiological study were consistent with the literature. Instead, pathogenicity was defined as “probable” if the above criteria were not met. For those with two allelic variants, the definite diagnosis required that compound heterozygous variants were confirmed. The final clinical-molecular diagnosis was made based on the genotype-phenotype correlation and was referred to Online Mendelian Inheritance in Man (OMIM) database.

Scenario 1: Patients with a positive WES result

Among these 17 patients with positive WES result (Table 1), most presented with a typical clinical phenotype. We made the diagnosis with genotype-phenotype correlation. In the group of definite diagnosis, two cases with atypical presentations (Cases 6 and 7). Case 6 was a male with exercise intolerance and high hyperCKemia. The work-up of metabolic myopathy and mitochondrial myopathy was normal. However, the patient had partial loss of dystrophin staining in muscle biopsy, supporting the diagnosis of dystrophinopathy. Case 7 had hand weakness as an initial symptom but later developed lower limb distal weakness, which is rarely reported in GNE myopathy.

There were 8 patients classified as members of the probable diagnosis group (myopathy: 5, neuropathy: 3). Three patients (CAPN3, TTN, and SCN4A) with atypical or uncommon clinical phenotypes could not meet the definite diagnosis criteria, and segregation analysis and molecular studies were needed. Four patients (FHL1, FBXO38, POLG, and OPA1) had variants that had been asserted as pathogenic or likely pathogenic in ClinVar database, and one patient (DYSF) need a segregation analysis to ascertain compound heterozygote.

Scenario 2: Patients with an inconclusive WES result but supporting clinical data

There were 9 patients showing inconclusive WES result but having supporting clinical data. Of note, 3 patients with myopathy who carried the new variants fulfilled the proposed definite diagnosis criteria. Their diagnosis was validated by (1) typical pathognomonic findings yielded from the clinical manifestations or muscle biopsy (e.g., distal myopathy and atrophic myocytes with desmin aggregates in DES-related myofibrillar myopathy 1; limb-girdle weakness with partial loss of dystrophin staining in muscle biopsy), or (2) further molecular examinations (e.g., mitochondria DNA sequencing and muscle electron transfer chain activity assay in TK2-related myopathy [22]).

There were another 6 patients classified in the probable diagnosis group (myopathy: 3, neuropathy: 3). Three patients (SLCO2A1, POLG, and SCO2) with representative clinical phenotypes could not meet the definite diagnosis criteria due to the lack of segregation analysis to ascertain compound heterozygote. Notably, in a patient with SLCO2A1-related hypertrophic osteoarthropathy we found a nonspecific myopathic change in the proximal muscles, which has not been previously reported. The patient with variant in TTN gene had early respiratory failure when she could still ambulate. However, we classified the case as a probable diagnosis because the variant was not located in the common region related to hereditary myopathy with early respiratory failure (HMERF). Although the patient had no necklace body in the muscle pathology, the abnormal desmin and dystrophin subsacrolemmal in light microscopy and Z-disk alterations in electron microscopy are compatible with the pathological diagnosis of myofibrillar myopathy, which has been reported in TTN-related HMERF [23, 24]. Two patients had variants in the NEFH gene. We made the diagnosis of brachial amyotrophic diplegia, primary lateral sclerosis, neuronopathy, distal hereditary motor neuropathy, and type IID based on the clinical presentation, electrophysiological studies, and WES results.

Scenario 3: Patients with negative WES result, but with final clinical diagnosis after further work-up

Notably, in the 15 patients without a molecular diagnosis from WES, four patients (26.7 %) had a diagnosis after further work-up. The details of these four patients are described below.

Case 27: Tumor-induced osteomalacia: A 32-year-old man presented with a 3-year history of progressive proximal muscle weakness, atrophy, and generalized back soreness. Electromyography performed at other hospitals showed myopathic changes with spontaneous activities and a right proximal tibial tumor. WES was performed but only found some benign variants. The clinical evaluation revealed a reduction in body height from 185 cm to 183 cm, kyphosis, and a palpable tender lump at the right proximal lateral tibia. Laboratory tests showed markedly reduced phosphate (1.3 mg/dL; normal range 2.5–5 mg/dL) and normal creatine kinase levels. The workup for the hypophosphatemia revealed normal calcium and iPTH, decreased 25-OH vitamin D (16 ng/mL; normal range 25–80 mg/mL), and elevated alkaline phosphatase (439 U/L; normal range 34–104 U/L). The urine analysis showed a markedly elevated urinary phosphate level (5332 mg/24 hrs). Furthermore, the bone scan showed a bone tumor with markedly increased osteoblastic activity at the right proximal femur. Combined with the above information, tumor-induced osteomalacia was highly suspected. Elevated serum fibroblast growth factor 23 (FGF23) confirmed the diagnosis. The pathology of the bone tumor was consistent with a phosphaturic mesenchymal tumor. The patient received tumor excision, his FGF23 was undetectable five hours after the operation, and his serum phosphate returned to normal three days after the surgery (Fig. 2). On follow-up, his bone pain and whole back soreness resolved. He could walk with a cane for more than 100 meters three months later and could stand without assistance for more than 10 minutes at the 5-month follow-up. (Supplementary Video)

Fig. 2

Muscle and bone histopathology in patient 28 with tumor-induced osteomalacia. The staining of hematoxylin and eosin (H&E) (A) and myosin adenosine triphosphatase at pH 4.3 (B) showed compacted muscle fibers with scattered atrophic fibers in the biopsy at left biceps brachii. The bone tumor pathology showed spindle tumor cells in myxohyalinized fibrotic stroma with eosinophilic matrix-like materials accompanying the bony components in H&E staining. (C&D) Focal bony destruction, woven bone formation and grungy calcification were also noted. The immunohistochemical staining of CD56 (E) and SATB2 (F) was strongly positive. Whole-body bone scintigraphy revealed a bone tumor with markedly increased osteoblastic activity at the right proximal femur. (G) Additionally, there were multiple posttraumatic changes in the thoracic cage, pubic bones, and bilateral femurs.

Case 28: Mitochondrial myopathy: A 41-year-old woman with no systemic disease visited our hospital for a one-month history of progressive proximal muscle weakness along with generalized myalgia. In addition, she lost eight kg in two months. She first received steroid therapy and her muscle strength improved in the first three months but later worsened even under treatment with prednisolone and azathioprine. Moreover, she developed obvious exercise intolerance. Her lactate level was 6.77 mmol/L at baseline (normal range 0.5–2.2 mmol/L). and increased to 11.96 mmol/L after exercise. Large amounts of lactate and pyruvate were noted in the urine as well as the presence of TCA cycle intermediates. The histopathology of the left biceps brachii revealed excessive lipid accumulation. (Fig. 3) Given the impression of metabolic myopathy, WES was ordered but yielded only benign variants. Based on the clinical information, we further arranged mitochondrial DNA (mtDNA) next-generation sequencing and a pathogenic variant (m.3243A>G) was found, confirming the diagnosis of mitochondrial myopathy.

Fig. 3

Muscle histopathology in patient 29 with mitochondrial myopathy. Hematoxylin and eosin staining showed variated muscle fiber sizes, degenerative and atrophic muscle fiber sizes, and nuclear internalization in the biopsy at the left biceps brachii. There were muscle fibers with variated cytoplasmic vacuoles. The vacuole was strongly positive for Oil Red O staining (B). In the staining of nicotinamide adenine dinucleotide-tetrazolium reductase (NADH-TR) (C) and succinate dehydrogenase (D), dark staining was noted in some muscle fibers (arrowhead). Vacuoles were also seen in the NADH-TR staining ( ^*).

Case 29: Microsatellite expansion disease: A 65-year-old man with congenital deafness developed unsteady gait at the age of 61. Bilateral upper limb clumsiness, dysmetria, and dysphagia appeared successively from the age of 63. With disease progression, he became unable to walk without assistance and was brought to our hospital for help. He reported a family history of gait problems in his mother and siblings. The neurological examination showed saccadic pursuit, dysphagia, decreased muscle strength in all four limbs, areflexia, positive Romberg test, and ataxic gait. The brain and spine MRI findings were normal. NCS revealed sensorimotor polyneuropathy. With multisystem involvement, WES was arranged but showed negative results. Notably, a heterozygous CAG repeat expansion (repeat number: 28 and 63) was noted in the ATX3 gene, and compatible with Machado-Joseph Disease.

Case 30: Mononeuropathy multiplex, vasculitis related: The 37-year-old woman presented with several episodes of mononeuropathy in all four limbs with or without complete recovery from the age of 17. The NCS disclosed absent or decreased CMAP in the bilateral median, ulnar, tibial, and right peroneal nerves. Additionally, there was absent or reduced sensory nerve action potential (SNAP) in the bilateral median, ulnar, sural, superficial peroneal, and left superficial radial nerves. The laboratory tests and CSF studies were within normal limits except for elevated rheumatoid factor. The combined biopsy at the right superficial peroneal nerve and right peroneal brevis muscle did not reveal evidence of vasculitis (Fig. 4). Under the impression of mononeuropathy multiplex with unknown cause, whole genome sequencing was arranged, yielding negative findings. Nevertheless, the patient-reported symptoms improved after receiving pentoxifylline and a short duration of oral steroids. Based on the clinical findings, the final diagnosis was vasculitis-related mononeuropathy multiplex.

Fig. 4

Combined muscle and nerve biopsy in patient 31. The staining of hematoxylin and eosin (H&E) (A) and nicotinamide adenine dinucleotide-tetrazolium reductase (B) showed grouped atrophy in the right peroneal brevis muscle. The biopsy in the right superficial peroneal nerve did not show thickening of small vessels, lymphocyte infiltration, red blood cell extravasation in H&E (C-E) and Masson’s trichrome staining (F).

Scenario 4: Patients with negative WES result and undetermined etiology in further work-ups

For the remaining 11 patients with a negative molecular diagnosis from WES, most (n = 7) had a clinical diagnosis of myopathy, while others (n = 4) were clinically diagnosed with neuropathy. The detailed clinical features of patients without definite diagnosis are presented in Table 4. Among these, some had negative findings from NGS. Two had suspected variants at first but were later excluded by segregation analysis. WGS may be helpful for these patients. For patients with unreported pathological variants of ANO5 and FKRP mutations, the definite diagnosis may be verified after segregation analysis and muscle biopsy with molecular studies. Currently, the patient is still hesitant about this plan. Some had suspected variants but without reported phenotypes, such as POLG-related Charcot-Marie-Tooth disease type 1 (CMT1) and ACAD9-related mitochondrial myopathy. Segregation analysis and further molecular diagnosis are needed.

Consent publication

All authors have read and agreed the contents of this manuscript submitted to Journal of Neuromuscular Diseases.