Patients with a phenotype consistent with facioscapulohumeral muscular dystrophy display genetic and epigenetic heterogeneity
- Sabrina Sacconi1,
- Pilar Camaño2,
- Jessica C de Greef3,4,
- Richard J L F Lemmers3,
- Leonardo Salviati5,
- Pascal Boileau6,
- Adolfo Lopez de Munain Arregui2,7,
- Silvère M van der Maarel3,
- Claude Desnuelle1
- 1Centre de référence des Maladies neuromusculaires and CNRS UMR6543, Nice University Hospital. Nice, France
- 2Department of Neurosciences, BioDonostia Health Research Institute, Donostia Hospital, 20014 San Sebastián, Spain and CIBERNED, Instituto de Salud Carlos III, Madrid, Spain
- 3Department of Human Genetics, Leiden University Medical Center, Leiden, the Netherlands
- 4Current address: Howard Hughes Medical Institute, Carver College of Medicine, University of Iowa, Iowa City, Iowa, USA
- 5Clinical Genetics Unit, Department of Pediatrics, University of Padova, Padova, Italy
- 6Department of Orthopaedic Surgery & Sports Traumatology, Archet 2 Hospital, Nice University Hospital, Nice, France
- 7Department of Neurology, Hospital Donostia, 20014 San Sebastián, Spain and Centro de Investiagación Biomédica en Red de Enfermedades Neurodegenerativas, CIBERNED, Instituto de Salud Carlos III, Madrid, Spain
- Correspondence to Dr Sabrina Sacconi, Centre de référence des maladies neuromusculaires, Archet 1 Hospital, 151 Route de Saint Antoine de Ginestière, 06000 Nice, France;
- Received 4 April 2011
- Revised 29 August 2011
- Accepted 30 August 2011
- Published Online First 7 October 2011
Objective To identify the genetic and epigenetic defects in patients presenting with a facioscapulohumeral (FSHD) clinical phenotype without D4Z4 contractions on chromosome 4q35 tested by linear gel electrophoresis and Southern blot analysis.
Design and patients The authors studied 16 patients displaying an FSHD-like phenotype, with normal cardiovascular and respiratory function, a myopathic pattern on electromyography, and a muscle biopsy being normal or displaying only mild and aspecific dystrophic changes. They sequenced the genes calpain 3 (CAPN3), valosin containing protein (VCP) and four-and-a-half LIM domains protein 1 (FHL1), and they analysed the D4Z4 repeat arrays by extensive genotyping and DNA methylation analysis.
Results The authors identified one patient carrying a complex rearrangement in the FSHD locus that masked the D4Z4 contraction associated with FSHD1 in standard genetic testing, one patient with somatic mosaicism for the D4Z4 4q35 contraction, six patients that were diagnosed as having FSHD2, four patients with CAPN3 mutations and two patients with a VCP mutation, No mutations were detected in FHL1, and in two patients, the authors could not identify the genetic defect.
Conclusions In patients presenting with an FSHD-like clinical phenotype with a negative molecular testing for FSHD, consider (1) detailed genetic testing including D4Z4 contraction of permissive hybrid D4Z4 repeat arrays, p13E-11 probe deletions, and D4Z4 hypomethylation in the absence of repeat contraction as observed in FSHD2; (2) mutations in CAPN3 even in the absence of protein deficiency on western blot analysis; and (3) VCP mutations even in the absence of cognitive impairment, Paget disease and typical inclusion in muscle biopsy.
- Neuromuscular disease
- muscle disease
- academic medicine
- molecular genetics
- immunology (including allergy)
- other endocrinology
- drugs: endocrine system
- metabolic disorders
- clinical genetics
Autosomal dominant facioscapulohumeral muscular dystrophy (FSHD) is characterised by facial and shoulder girdle muscle weakness, with progression of the disease to anterior foreleg, abdominal and pelvic girdle muscles.
Facial involvement frequently includes orbicularis oculi and oris muscles, resulting in impaired palpebral occlusion and transverse smile and determining the typical facial appearance of these patients.1 Asymmetry of muscle involvement is considered as a typical feature, but is not always present.2
The genetic lesions in the majority of FSHD cases are linked to the subtelomeric region of chromosome 4q (FSHD1).3 This region contains a polymorphic repeat array consisting of 3.3 kb repeated D4Z4 units which can vary in the general population between 11 and 100 units. In patients with FSHD1, at least one 4q35 residing D4Z4 repeat array is contracted to 1–10 units. Standard molecular diagnosis is based on Southern blot analysis of restriction enzyme-digested DNA fragments after linear gel electrophoresis4 using the probe p13E-11, which hybridises proximally to the D4Z4 region.3 This technique allows for a correct molecular diagnosis in most of the patients displaying the classical clinical picture of FSHD; nevertheless, in a limited number of patients, it yields false-positive or false-negative results.5 Indeed, two distinct 4q chromosomes (4qA and 4qB) were identified, differing essentially for the presence of a 0.3 kb pLAM fragment and a 6.2 kb β satellite DNA fragment immediately distal to the D4Z4 locus on 4qA chromosomes. Only D4Z4 repeat contractions on 4qA alleles are pathogenic.6 7 In addition, as a result of an ancient duplication, the subtelomere of chromosome 10q also contains a repeat array highly homologous to 4q D4Z4, but repeat contractions on this chromosome are not associated with FSHD.4 The pLAM fragment contains the 3′ untranslated region and the polyadenylation (poly(A)) signal of the DUX4 gene which is located in each D4Z4 unit.8 9 Contractions of D4Z4 in FSHD1 lead to local chromatin relaxation, with loss of DNA methylation, and transcriptional derepression of the DUX4 gene.9–12
Multiple haplotypes of chromosome 4q have been distinguished on the basis of sequence variations in the FSHD locus, yet contractions in only a few of these cause FSHD. More specifically, contractions of D4Z4 on 4A166 chromosomes appear to be non-pathogenic, or at least much less pathogenic, as evidenced by the identification of multiple healthy family members in two Dutch families carrying contracted D4Z4 repeat arrays on a 4A166 chromosome.13 A detailed description of the structure of these alleles has been reported elsewhere.9 13 14
Consequently, to avoid false positives, exact determination of the genetic background of chromosome 4q is necessary to confirm FSHD in sporadic cases and in patients presenting with an atypical clinical phenotype.5 13
Some patients with FSHD1 may harbour a contraction of repeats with non-canonical D4Z4 composition consisting of 4q-type and 10q-type D4Z4 units.14 These arrays may appear as chromosome-10q-derived repeats and complicate FSHD molecular diagnosis, generating false-negative results.15 Alternatively, patients may carry a proximal deletion including the p13E-11 probe region that cannot be detected by the standard approach.16
A specific subgroup of patients with FSHD presents with chromatin relaxation and DNA hypomethylation on non-contracted D4Z4 repeat arrays in the presence of at least one permissive chromosome. These patients have been referred to as FSHD2 patients and no causative mutation in a specific gene has been found yet.17–19
Mutations in other genes may also mimic FSHD. For example, calpain 3 (CAPN3) mutations have been associated with autosomal recessive limb girdle muscular dystrophy type 2A (LGMD2A), which, in a subgroup of patients, may present with scapulohumeral muscle weakness resembling FSHD, a milder phenotype compared to the classic form, with later onset and no contractures.20 Diagnosis of LGMD2A is currently based on western blot analysis of a muscle biopsy, but genetically confirmed patients with normal protein expression levels have been reported,21 22 and, interestingly, some of them displayed a scapulohumeral phenotype.20
Mutations in valosin containing protein (VCP) have been associated with an autosomal dominant progressive disorder comprising inclusion body myopathy, Paget's disease of bone and frontotemporal dementia. VCP patients may display isolated muscular weakness with scapuloperoneal distribution resembling FSHD.23 Finally, X linked dominant scapuloperoneal syndrome has been reported in association with mutations encoding four-and-a-half LIM protein 1 (FHL1).24
Patients and methods
Inclusion criteria for the study were an FSHD clinical phenotype defined as the presence of three or more of the following features: (1) evidence of autosomal dominant inheritance and/or weakness in (2) facial muscles, (3) shoulder girdle muscles, (4) anterior foreleg muscles and (5) asymmetrical muscle involvement. In addition, all patients had (1) normal serum or plasma biochemical parameters (with the exclusion of creatine kinase (CK)) including erythrocyte sedimentation rate, electrolytes, thyroid and parathyroid hormones; (2) normal electrocardiogram, echocardiography and respiratory pulmonary function; (3) electromyography (EMG) with myogenic pattern in affected territories without evidence of myotonic discharges and (4) a muscle biopsy showing no abnormalities or only mild and unspecific dystrophic changes. Muscle samples were processed using standard procedures.25
For each patient included in this study, data concerning age, sex, family history, distribution of weakness and CK blood levels were collected (table 1) after informed consent. None of these patients displayed contractures or involvement of ocular muscles.
Clinical data collection, extensive D4Z4 genotyping and D4Z4 methylation analysis were also carried out in all family members of patients with FSHD2 that were available after informed consent.
Single sequence length polymorphism, repeat length and distal variation determination were performed as described14 (see also the Fields Center for FSHD Research website: http://www.urmc.rochester.edu/fields-center/).
Sixteen patients (P1–16) fitted the inclusion criteria described above: 10 men and 6 women. The mean age at clinical evaluation was 48.5±7.2 years, and the mean age at onset of the disease reported by patients was 29±6.6 years. Clinical characteristics of the patients are reported in table 1. Genomic DNA of all patients was subjected to CAPN3, VCP and FHL1 mutation analysis by direct exon sequencing, to extensive D4Z4 genotyping by pulse field gel electrophoresis (PFGE) analysis and to DNA methylation studies of the D4Z4 repeat by methylation-sensitive Southern blot analysis.
Patient 1: somatic mosaicism for D4Z4 contraction: FSHD1
Patient P1 is a sporadic case and was found by PFGE analysis to be mosaic for a contracted D4Z4 repeat on a 4A161 allele of 11 kb in ∼55% of her blood cells, corresponding to one D4Z4 repeat unit. Reported disease onset was at age 35 years: she showed a typical FSHD clinical phenotype including asymmetric facial weakness, shoulder girdle muscle involvement, and abdominal and anterior foreleg muscle weakness. The milder phenotype displayed by P1, still ambulant at the age of 44 years, which was not expected based on the small residual repeat size (one repeat unit), may be explained by the mosaic nature of the contraction.
Patient 2: hybrid D4Z4 repeat: FSHD1
This patient, a 48-year-old woman, presented at age 5 years with a typical FSHD phenotype including facial, scapulohumeral and anterior forelegs asymmetric muscular weakness. At age 45 years, abdominal and pelvic girdle muscles were also involved.
Her 27-year-old daughter was also affected, presenting with mild facial and shoulder girdle involvement and with right dorsiflexor muscle weakness. She carried the same genetic defect, while the two younger twin sisters did not carry the pathogenic allele and were asymptomatic.
Facial weakness in both patients consisted of asymmetric involvement of orbicularis oculi and orbicularis oris (supplementary figure 1).
Genotype studies of this family have been described in a previous report (family 3 of Lemmers et al).9 Briefly, the pathogenic chromosome is the result of a meiotic exchange between chromosomes 10q and 4q creating a contracted hybrid D4Z4 repeat array on a permissive chromosomal background.
Interestingly, all carriers of this contracted hybrid D4Z4 repeat are less severely affected than what was expected based on the residual number of D4Z4 units, which is one. We suggest that the composition of the array, which mainly consists of non-permissive 10q sequences, may play an important role in determining the milder phenotype. On the other hand, less severe FSHD phenotypes have been often described in females.30
Patients 3, 4, 5, 6, 7 and 8: FSHD2 patients
These patients were all sporadic cases, displaying a typical FSHD phenotype. Interestingly, most of them experienced pain and fatigue, symptoms frequently reported by FSHD1 patients,31 and one of them (P3) had sensorineural hearing loss, a feature described in 25%–65% of FSHD1 patients.32 CK values were normal to twofold elevated, and muscle biopsy showed only mild dystrophic changes. All these patients carried a 4A161 allele of 45–95 kb and marked hypomethylation of the D4Z4 locus, typical of FSHD2 (table 2).
Pedigrees of P3, P4, P6, P7 and P8 are reported in supplementary figure 2, while the pedigree of P5 has already been reported and discussed in detail (family 6 in de Greef et al19).
Family members of these patients available have also been clinically characterised, and D4Z4 genotyping and methylation studies were performed.
The mothers of P3, P6 and P7 and the father of P8 carried the same 4A161 allele as their affected offspring, but they were not hypomethylated and were asymptomatic (supplementary figure 1A, C, D, E). Interestingly, the 26-year-old daughter of P4 displayed significant hypomethylation, but she had D4Z4 repeats on a non-permissive chromosome 4 background and is asymptomatic (supplementary figure 1B). This suggests that the hypomethylation determinant segregates independently from D4Z4 at 4q35 and therefore is likely to be at a different locus on the genome.
Patients 9 to 12: LGMD2A
Patients P9–P12 were sporadic cases, presenting with symmetric facial and with shoulder girdle, pelvic, and anterior foreleg muscle weakness. All carried pathogenic mutations in CAPN3 (table 3).28 29 33 In these four patients, western blots for calpain 3 protein in muscle were normal, and muscle biopsies were characterised by mild dystrophic features.
Mean CK values were three and four times the upper limit of normal values in P9 and P10, which are higher than the mean CK values classically found in patients with FSHD, while P11 and P12 had twice the upper limit of normal CK values. Facial involvement was present in all patients and consisted essentially of a hypotonic and hypokinetic face without involvement of orbicularis oculi muscles. Indeed, patients were unable to whistle and to blow a candle, but they could properly close their eyes and mimic a kiss.
Patients 13 and 14: VCP mutations
Patient P13 had clinical evidence of autosomal dominant inheritance since the father was diagnosed as having a myopathy of unknown origin. He displayed a scapuloperoneal phenotype with a certain degree of asymmetry in the legs. Facial involvement was not present, and CK values were three times the upper limit of normal values. The muscle biopsy was mildly dystrophic without inclusion bodies.
P14 was adopted and displayed facial involvement, shoulder and pelvic girdle weakness, and anterior foreleg symmetric muscle weakness. CK values were fourfold increased. Interestingly, facial involvement presented with difficulties in closing the eyes and protruding the lips, as are typically seen in patients with FSHD.
The muscle biopsies of P13 and P14 showed mild dystrophic changes and no inclusion bodies. Both carried a mutation in the VCP gene reported in table 3, previously described as pathogenic.23 After the molecular diagnosis, they underwent an extensive neuropsychological evaluation that was normal in P14 while revealing a mild dysexecutive syndrome in P13, characterised by low scores in short-term memory scales. Brain MRI, serum alkaline phosphatase activity and skeletal x-ray (pelvis, hip, femur, humerus, vertebral bodies and skull) were normal in both patients.
Patients 15 and 16
P15 and P16 are sporadic cases with a typical FSHD phenotype, including facial weakness, and no peculiar features in EMG or in muscle biopsy that may suggest a specific form of muscular dystrophy. CK values were four and six times the upper limit of normal values. No mutation was found in all genes sequenced, and the genotype and methylation study of the D4Z4 locus did not show evidence for FSHD or FSHD2. This may suggest the presence of a mutation in a gene mimicking an FSHD phenotype, which has not yet been identified.
The clinical diagnosis of FSHD is frequently not an issue for a neurologist expert in neuromuscular disorders due to the typical pattern of muscle involvement. Nevertheless, the classical genetic approach based on identification of 4q35 D4Z4 contracted allele after EcoRI and EcoRI/BlnI digestion, linear gel electrophoresis, Southern blotting and p13E-11 hybridisation fails to offer the molecular confirmation in about 5%–10% of cases.14
The present study was conducted to guide the clinician in the choice of additional genetic or epigenetic tests to confirm the clinical diagnosis since several conditions can partially mimic the clinical phenotype of FSHD.
In clinical practice, when standard genetic testing does not confirm the clinical suspicion of FSHD1, EMG is required to confirm the myogenic pattern of muscle weakness and to rule out neurogenic involvement. For example, patients harbouring deletions of PMP2234 or mutations in the TRPV4 gene,35 both transmitted as autosomal dominant traits, may present with a scapuloperoneal distribution of muscular involvement of neurogenic origin.
Other inherited myopathies may present with a clinical phenotype resembling FSHD: some limb girdle muscular dystrophies (LGMD),36 glycogenosis type II (Pompe disease)23 and glycogenosis type V (McArdle's disease),37 myosin storage myopathies,38 desminopathies,39 other myofibrillar myopathies24 40 and mitochondrial myopathies.41
In most of the cases, the age of onset, the disease progression rate, the pattern of inheritance, the presence of respiratory muscle or cardiac involvement, the absence of facial weakness, the presence of contractures or ocular involvement may aid the clinician in orientating the diagnosis towards these diseases.
In all these cases, muscle biopsy is mandatory to confirm the clinical suspicion. Indeed, histological, immunohistochemical and western blot studies on muscle samples are generally adequate to distinguish among different LGMDs due to lack of sarcolemmal protein or to identify storage, vacuolar, mitochondrial, congenital or myofibrillar myopathies.
If EMG and muscle biopsy studies are not conclusive, the use of detailed PFGE-based D4Z4 genotyping on agarose-embedded DNA plugs42 can be used to identify patients with FSHD1 that could have been missed by the less detailed standard molecular diagnosis.
In our study, we found two such patients among the 16 patients included in the study: P1 and P2. P1 was mosaic for the D4Z4 contraction. This condition results from a mitotic rearrangement of D4Z4 probably in early embryogenesis and is quite frequent in de novo cases,43 44 and mosaicism often goes undetected.45 Their clinical phenotype appears to be milder compared to individuals carrying the same residual repeat size in all of their cells. In P2, the pathogenic hybrid allele was identified by PFGE, and the distal D4Z4 fragment was analysed by direct sequencing to confirm the pathogenicity.9
In all cases without a diagnosis and with a permissive chromosome, FSHD2 should be suspected.10 17–19 Indeed, patients with FSHD2 display (1) a clinical phenotype identical to FSHD1 patients, (2) normal-sized repeats with at least one permissive chromosome (contrary to FSHD1 where the repeat is contracted on a permissive chromosome) and (3) reduced DNA methylation of the D4Z4 repeats on chromosomes 4q and 10q, as measured on the methylation-sensitive restriction sites FseI and BsaAI.
In our series, six patients (P3–P8) were found to meet these criteria. All of them were sporadic cases, and we studied the segregation of the permissive allele and the DNA methylation in all members available for the six families, confirming that the FSHD clinical phenotype appears only in patients displaying hypomethylation on a permissive 4A161 background. Finally, when EMG studies, muscle biopsy studies and extensive D4Z4 repeat arrays genotyping and DNA methylation analysis are negative, we suggest reconsidering the possibility of an FSHD phenocopy.
In our series of patients, six out of the 13 patients (P9–P14) carried mutations in genes not related to FSHD, that is, CAPN3 and VCP. These genetic defects had gone unnoticed by standard histological, immunohistochemical and western blot studies on muscle biopsy.
Actually, approximately 20% of CAPN3 mutations do not cause a reduction of the protein steady-state levels, but affect its autocatalytic activity. A specific test is available to measure this activity in muscle, avoiding these misdiagnoses.22 Unfortunately, this test is done only in a restricted number of laboratories and, in the case of the present study, was not performed because no muscle sample was left. Furthermore, it has been reported that patients harbouring VCP mutation may display only minor and non-specific histopathological changes on muscle biopsies.14
Accordingly, in P13 and P14, carrying a mutation in the VCP gene, no sign of Paget's disease and no characteristic rimmed vacuoles on muscle biopsy were found. Nevertheless, after the molecular diagnosis, P14 underwent a neuropsychological test demonstrating the presence of a mild dysexecutive syndrome. Alkaline phosphatase activity measurement, standard bone x-ray (pelvis, hip, femur, humerus, vertebral bodies and skull) and adapted neuropsychological tests may help in detecting asymptomatic Paget's disease and frontotemporal dementia that are usually associated with an inclusion body myopathy and mutations in the VCP gene.27
Interestingly, CK levels in four out of the six patients displaying a phenotype resembling FSHD but carrying mutations in other genes were more elevated than in classical FSHD1 patients, and five out of six patients displayed no asymmetric muscular involvement.
Concerning facial weakness, P9, P10, P11 and P12 (LGMD2A) and P14 (VCP) showed transverse smile and difficulty to blow, as did P4 and P6 (FSHD2), while P2 (FSHD1), P3, P5, P7 and P8 (FSHD2) showed also weakness of orbicularis oculi (supplementary figure 1).
Altogether, these clinical features, if present, may prompt clinicians to sequence CAPN3 and VCP gene before undergoing more complex epigenetic studies.
Mutations in FHL1 have been described to cause a scapuloperoneal phenotype mimicking FSHD, inherited in an X dominant fashion.24 We could not find FHL1 mutations in our patients, possibly because they are less frequently associated with an FSHD-like phenotype. Nevertheless, we cannot exclude that mutation in this gene may be associated to FSHD-like phenotype due to the small number of patients included in the present study.
In two out of 16 patients (P15, P16) included in this study, no genetic or epigenetic defect could be found. In these two sporadic patients, CK values were elevated four to six times the upper limit of normal values, which may suggest the presence of other loci determining an FSHD-like phenotype. In the absence of an identified genetic or epigenetic defect, FSHD-like patients should undergo a complete clinical workup in order to identify additional signs that may be useful to orientate the diagnosis. Annual follow-up is also recommended to monitor the pattern of muscle weakness progression and, possibly, the appearance of signs of symptoms evocative of a specific muscular dystrophy. The possibility of a second muscle biopsy must be carefully discussed with the patient in case of worsening of clinical conditions.
In conclusion, in patients with an FSHD phenotype and no D4Z4 contraction on chromosome 4q35 detected by standard techniques, it is important to consider these possibilities: (1) a false-negative diagnosis, (2) FSHD2 and (3) a mutation in another gene resembling FSHD (especially if CK levels are consistently elevated and there is evidence of symmetric muscle involvement) (see the diagnostic flow chart in figure 1). Extensive D4Z4 genotyping by PFGE analysis and methylation studies of the D4Z4 repeat, which are currently available only in selected research laboratories, are required to diagnose the first two conditions.
However, novel diagnostic approaches are currently being developed to simplify the diagnosis of FSHD and other muscle disorders. Molecular combing is a promising new strategy which may detect mosaicism and complex alleles, and is apparently simpler and faster than Southern blot and allows 4q allele determination in a single step procedure.46 Furthermore, next-generation sequencing technology will allow study of large panels of genes for a fraction of the cost (and of the time) required by classic Sanger sequencing and will greatly enhance our diagnostic yield in patients with FSHD-like phenotypes.
We acknowledge the patients and their families, and the French Association against Myopathies (AFM) for its help in referring patients.
Funding Pilar Camaño and Adolfo Lopez de Munain Arregui are funded by the Centro Investigación Biomédica en Red para Enfermedades Neurodegenerativas (CIBERNED), the Basque government (Fellowship grant, No. 2008111011), the Instituto Carlos III, ILUNDAIN Fundazioa, the Prinses Beatrix Fonds, the Fields Center for FSHD and Neuromuscular Diseases, and the National Institutes of Health (P01NS069539, AR059966) to take care of clinical evaluation and molecular studies (VCP and CAPN3 gene).
Competing interests None.
Ethics approval This is a retrospective study which involved the standard diagnostic procedures for patients with myopathy. All procedures (biopsies, genetic tests, photographs, etc) were performed with the written informed consent of the patients.
Provenance and peer review Not commissioned; externally peer reviewed.