Exome sequencing identifies novel dysferlin mutation in a family with pauci-symptomatic heterozygous carriers

Patients and clinical evaluations

The study and subject consent forms were approved by the University of Cape Town Health Sciences faculty human ethics research committee (REF 552/2013) and the study carried out in accordance with the approved guidelines and regulations. We studied two generations of a South African family of mixed genetic ancestry that likely includes ancestors from Africa, Europe as well as Madagascar and Java as previously described [12]. Family members of G1 and G2 had been attending the adult and paediatric neurology clinics attached to the University of Cape Town for 25 years. After signed informed consent, blood was obtained for WES from G1 individuals, three affected (I-2, I-4 and I-5), one unaffected sibling (I-1) and one unrelated unaffected family member (I-3) as controls, and two members from G2 with neuromuscular symptoms and areflexia, II-2 and II-5. Structured folder reviews were performed to obtain previous examination details, laboratory data including those of muscle biopsies and clinical electrophysiological studies.

Whole exome capture, sequencing and variant calling

Paired-end exome sequencing was performed at 50× coverage by CLIA accredited Otogenetics Corporation, Norcross, GA, USA using the Agilent SureSelect Human V5 + UTR capture kit and the Illumina HiSeq2000/2500 platform. After quality control, reads for each patient were aligned to the hg19 human reference genome using NOVOALIGN [13], PCR duplicates were removed using Picard [14], followed by indel realignment and base quality score recalibration, variant calling and quality evaluation using the Genome Analysis ToolKit [15] version 3.6 to produce a high-confidence set of variants for each sample.

Identification of likely function-impacting candidate variants

Variants were annotated using ANNOVAR [16] and were filtered based on both autosomal recessive and dominant inheritance models. For the recessive model, we identified variants that were homozygous in the three affected G1 siblings, wild-type in their unaffected sibling, and wild-type or heterozygous in the unrelated unaffected family member and the G2 offspring. For the autosomal dominant model, we identified variants that were heterozygous in all the affected/pauci-symptomatic family members, and where both related and unrelated controls were wild-type. For initial filtering, variants with a minor allele frequency > 1% in public databases, namely the 1000 Genomes Project [17], the NHLBI-ESP 6500 Exome Sequencing Project [18] and the Exome Aggregation Consortium [19] databases were filtered out, as were those present in the currently unreleased Southern African Human Genome Program [20] variant dataset. Nonsense, frameshift and splicing variants were automatically selected as preliminary candidates, while missense variants were further evaluated if they passed the recommended deleteriousness score thresholds for any one of FATHMM [21] or the MetaSVM and MetaLR ensemble prediction methods [22].

Clinical findings

Three affected individuals from G1 had developed progressive muscle weakness in their twenties, two of whom also had prominent exercise-induced myalgia, and five individuals from the second generation developed neuromuscular symptoms in adolescence (Fig. 1). Two individuals from G1 (I-4 and I-5) presented with the distal Miyoshi muscle dystrophy phenotype and one (I-2) with a “proximodistal” phenotype. Two cases from G2 had areflexia and marginally raised CK (II-2 and II-5), and three (II-1, II-6, II-7 and II-8) had normal reflexes. Individuals I-1, II-3, and II-4 had no neuromuscular symptoms or signs. The offspring of 1–1 were asymptomatic adults, and those of I-4 had not reached adolescence and were not examined.

The index case (I-4), first examined at age 26, complained of two years of progressive thinning of the distal legs and quadriceps, an inability to stand on his toes, and myalgia in his legs aggravated by exercise. Later, he noticed weakness in the hands. The clinical presentation was a distal, posterior compartment muscular dystrophy with markedly raised creatine kinase (CK) levels (> 47× upper limit of normal (ULN)). However, the tendon reflexes were either absent (legs) or reduced (arms), and four years later all the reflexes were absent. Electrophysiology showed normal nerve conductions, and myopathic features on needle electromyography (EMG). Electrocardiography was normal. A deltoid muscle biopsy confirmed a dystrophic process without inflammatory infiltrates; no special staining was available. At that time, his two asymptomatic sisters (one was Case I-5, aged 21, see below) and mother were noted to have normal CK levels (CK ≈ 61 IU/L; N = 26–140 IU/L), although his father, aged 51, complained of muscle cramps and had a slightly raised CK level (249 IU/L) which was 1.5× above the upper level of normal expected for age and sex. Approximately 15 years after symptom onset, I-4 required bilateral crutches to mobilize, and 10 years later became wheelchair bound. The pattern of weakness had progressed to severe limb girdle and distal weakness. Although sensation was previously recorded as intact, the last examination at age 49 showed evidence of a mild sensory stocking neuropathy.

Case I-2 presented at age 29 with a history of progressive leg weakness since mid-twenties, noting difficulty climbing stairs, getting up from chairs and exercise-induced muscle cramps, especially in his calves. His examination showed wasting of the biceps and distal legs, mild proximo-distal posterior leg weakness, and reduced/absent reflexes. Later he exhibited a waddling gait. The CK level was 20× ULN. A muscle biopsy at age 40 showed features of muscle dystrophy and immunohistochemistry showed absence of dysferlin in the presence of positive merosin, emerin, caveolin, dystrophin and sarcoglycan staining. Fifteen years later he required crutches to mobilize.

Case I-5 presented at age 34 with increasing difficulty in walking, climbing stairs, rising from a seated position, and general muscle fatigue since her late twenties. The arms showed tapering distally and the posterior compartment of the legs, marked wasting. Her tendon reflexes were globally depressed and there was marked weakness of the posterior compartment leg muscles. Her CK level was > 30× ULN. A muscle biopsy of the left biceps showed similar results to her brother except that dysferlin staining was initially present and dystrophin was absent, but the positive control (spectrin) showed partial staining. Her disability increased substantially and at age 45 years she was largely confined to a wheelchair. A repeat biopsy of the right biceps showed absent dysferlin staining.

II-2 was examined at age 18 years. He had normal early motor development but was noted to fall more than usual as a child whilst running. Muscle cramps and stiffness, especially with physical activity, was noted during early adolescence. At age 15 he had increasing difficulty with riding his bicycle, climbing stairs as well as gait instability and stopped playing sport. He fell frequently. The neurological examination showed mild wasting of the biceps despite well-developed muscles elsewhere. His tendon jerks were absent. Power testing was normal, but he had a mild waddling gait. Sensation and coordination testing were normal. Clinical electrophysiology was refused. The CK level was at the upper the limit of normal (ULN) for his age (218 IU/L).

II-5 had experienced muscle pain, and episodes of cramp and stiffness lasting several hours to 1–2 days since the age of 12–13 years. These symptoms were, and still are, aggravated by physical activity. Since his early twenties he has also noticed increasing clumsy ankles and occasional give-in weakness of the legs. Examination at age 25 showed floppy ankles but no obvious wasting in his hands or feet. His tendon reflexes were globally absent and he had mild weakness of toe flexors but not of the plantar flexors. The remaining motor, sensory (all modalities) and coordination systems were normal. The CK level was slightly elevated (193 IU/L; 1.1× ULN for age and sex). Nerve conduction studies were normal. EMG of the medial head of gastrocnemius showed no spontaneous activity and normal motor units.

Cases II-1, II-6, II-7 and II-8 had been experiencing neuromuscular symptoms since the ages of 12 to 13 years, mainly myalgia, stiffness and/or muscle cramps. These symptoms occurred either in the hands or legs and were frequently provoked by mild physical activity including writing with a pen or walking upstairs, respectively.

Exome sequencing and filtering for candidate deleterious variants

WES was performed on I-2, I-4 and I-5 from G1, one unaffected sibling (I-1) and one unrelated unaffected family member (I-3) as controls, and on two members from G2 with neuromuscular symptoms and areflexia, II-2 and II-5. Approximately 50,000 variants were identified for each individual. Six variants segregated with all three symptomatic G1 individuals using the recessive model, of which three were in exonic regions, and one each in a 3′-untranslated region (UTR), intronic and intergenic regions, respectively. A novel nonsense C > G mutation located at position 4299 in exon 39 (NCBI RefSeq: NM_003494) of the dysferlin gene (DYSF) was identified as responsible for the family’s muscular dystrophy after filtering out common variants and those not predicted to impact protein function. Sanger sequencing confirmed the mutation to be homozygous in affected individuals of G1, and carrier status in I-1, II-2 and II-5. The p.Tyr1433Ter (NP_003485.1) mutation results in the loss of the sixth (C2E) and seventh (C2F) domains of the dysferlin protein (Fig. 2).

No other potentially deleterious dysferlin variant was found in the exomes of the two pauci-symptomatic G2 DYSF-mutant carriers. As they also presented with neuromuscular symptoms and areflexia in absence of distinct myopathy, we further filtered for rare or novel variants predicted to impact protein function that segregated exclusively with symptomatic individuals in both G1 and G2. Only two novel heterozygous missense mutations were identified in the LRP2 (low density lipoprotein-related protein 2) and GXYLT1 (glucoside xylosyltransferase 1) genes, which have no known roles in neuromuscular disease.