Objective Several families with characteristic features of hereditary myopathy with early respiratory failure (HMERF) have remained without genetic cause. This international study was initiated to clarify epidemiology and the genetic underlying cause in these families, and to characterise the phenotype in our large cohort.
Methods DNA samples of all currently known families with HMERF without molecular genetic cause were obtained from 12 families in seven different countries. Clinical, histopathological and muscle imaging data were collected and five biopsy samples made available for further immunohistochemical studies. Genotyping, exome sequencing and Sanger sequencing were used to identify and confirm sequence variations.
Results All patients with clinical diagnosis of HMERF were genetically solved by five different titin mutations identified. One mutation has been reported while four are novel, all located exclusively in the FN3 119 domain (A150) of A-band titin. One of the new mutations showed semirecessive inheritance pattern with subclinical myopathy in the heterozygous parents. Typical clinical features were respiratory failure at mid-adulthood in an ambulant patient with very variable degree of muscle weakness. Cytoplasmic bodies were retrospectively observed in all muscle biopsy samples and these were reactive for myofibrillar proteins but not for titin.
Conclusions We report an extensive collection of families with HMERF with five different mutations in exon 343 of TTN, which establishes this exon as the primary target for molecular diagnosis of HMERF. Our relatively large number of new families and mutations directly implies that HMERF is not extremely rare, not restricted to Northern Europe and should be considered in undetermined myogenic respiratory failure.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Hereditary myopathy with early respiratory failure (HMERF, OMIM #603689) was described as an autosomal dominant disease characterised by adult onset proximal and/or distal myopathy with respiratory failure typically in ambulant patients. HMERF disease has been associated with two different titin mutations.1–4 Muscle histopathological features include a combination of cytoplasmic bodies (CBs) and rimmed vacuoles.3 A diagnostic pattern of fatty degenerative changes in lower limb muscles on MRI has been identified, showing marked involvement of semitendinosus, obturator, sartorius, gracilis and iliopsoas muscles.3–6 Several families and patients with clinical, morphological and imaging features compatible with HMERF have remained without molecular genetic cause, including two of the originally reported families1 and two previously described separate families.6 ,7 In the current large international study we identified the recently reported titin A-band mutation3 ,4 in six families from various ethnic backgrounds and, moreover, four novel titin mutations in the same A-band domain in six other families indicating that the disease is not extremely rare and maybe underdiagnosed.
Mutations in TTN gene, encoding the giant muscle protein titin, are known to cause several different skeletal and/or cardiac myopathies.8–11 A dominant mutation in the kinase domain of M-line titin leading to HMERF was first described in three families from Sweden.2 Two recent reports revealed a novel titin A-band mutation in three Swedish and three British families with HMERF.3 ,4 Based on our large international collection of families and new mutations, the range of epidemiological, clinical and mutational spectrum of HMERF disease expands considerably and includes unexpected semirecessive inheritance with one of the mutations.
Study protocol and patients
Patients belonged to 12 unrelated families: one French (A), one Finnish (B), two Swedish (C, D), two British (E, I), two Italian (F, J), one Argentinian (G), one German (H) and two French with Portuguese ancestry (K, L) (figure 1). Families A–F and H–J correspond to the original Caucasian populations in these countries, whereas the Argentinian family has a Caucasian, probably Italian background. The collection contains all families with HMERF known to the authors without established molecular genetic cause. Among the 31 affected patients, 16 were male and 15 female with a mean age at examination of 49 years (range 29–72 years). All patients had been clinically examined by neurologists including muscle strength evaluation (the Medical Research Council Scale, MRC) and medical and family histories. Pulmonary function tests and echocardiography were performed, as well as electrophysiological examinations consisting of nerve conduction studies and needle electromyogram (EMG), creatine kinase measurement and muscle imaging by CT or MRI. The diagnosis of HMERF was based on clinical symptoms of respiratory insufficiency with muscle weakness and/or presence of CBs in muscle biopsy and a typical pattern of muscle involvement on MRI as described previously.3 ,4 ,6 ,7 DNA samples of 31 affected patients, as well as 26 healthy family members were provided by the clinicians in the different countries.
Muscle biopsies were obtained from proximal muscle (deltoid or thigh muscle, 20 samples). They were snap frozen and 8–10 µm sections were cut and examined using standard histochemical stainings including H&E, Gomori trichrome, reduced nicotinamide adenine dinucleotide-tetrazolium reductase and ATPase at pH 10.4, pH 4.6, and combined succinate dehydrogenase and cytochrome oxidase. Four samples were stained for actin with rhodamine conjugated phalloidin. Sections were also immunostained for different myogenic antigens including myosin heavy chain (MyHC) isoforms (fetal, neonatal, slow and fast MyHC, MHC (major histocompatibility complex) class I). Five samples were available for additional immunohistochemistry.
Fluorescently labelled polymorphic microsatellite markers spanning a region of 7 Mb in the TTN locus were used for genotyping the families. Markers used were D2S2314, D2S1244, D2S138, D2S148, D2S2173, D2S300, D2S385, D2S324, D2S2978, D2S2261, D2S384, D2S364, and D2S350. Genotyping was performed using ABI3730xl DNA Analyzer and GeneMapper V.4.0 software (Applied Biosystems).
Two affected (II-3 and III-1) members and one healthy (II-1) member of French family A (figure 1) were exome sequenced at Axeq/Macrogen laboratory in South Korea. The capture method used was Illumina TruSeq Exome Enrichment. Captured DNA fragments were sequenced on an Illumina HiSeq2000 platform using 100 bp paired-end reads. Sequence reads were aligned to the human reference genome (UCSC hg19) using the Burrows-Wheeler Aligner.12 Variant calling was made with GATK.13 Variant quality/control data filtering was performed using the analysis and visualisation programme RikuRator (unpublished), created by Riku Katainen from Lauri Aaltonen's group at the University of Helsinki. To call a variant, the coverage was required to be at least two reads and the mutated allele to be present in at least 20% of the reads. Only those variants that were shared by both the affected were included and filtered against dbSNP132 and one healthy member of the family.
Mutations were confirmed by Sanger sequencing. Sequencing primers were obtained from Genethon and are available on request. PCR was performed with DreamTaq DNA Polymerase according to standard protocol (Fermentas). PCR products were sequenced on an ABI3730xl DNA Analyzer (Applied Biosystems), using the Big-Dye Terminator V.3.1 kit and analysed with Sequencher V.5.0 software (Gene Codes Corporation).
Clinical characteristics of the patients
Clinical data of 31 patients with HMERF are presented in the online supplementary table and pedigrees of the 12 families in figure 1. The presenting symptom was either lower limb weakness (14/22) or respiratory failure (8/22) with a mean age of onset of 36.6 years (range 16–53 years). In family D four patients reportedly already had muscle weakness in childhood but no results of examinations performed in childhood were available for confirmation. Muscle weakness was progressive and usually symmetrical despite occasional asymmetry on imaging (table 1). Disease duration at the time of the latest examination was on average 13 years (range 1–32 years) and the most typical symptoms consisted of distal and proximal lower limb weakness and respiratory insufficiency that needed invasive or non-invasive ventilation. Neck flexor, abdominal and ankle dorsiflexion weakness was marked. There was no upper limb weakness at onset. In the later course of the disease proximal and distal weakness in the upper extremities was observed in 19 patients. The severity of muscle weakness and its rate of progression varied from mild (no limb weakness in four patients) to loss of ambulation at age 36 years. One member of family A did not have any weakness in limbs or respiratory muscles at age 38 but did show mild pathognomonic findings in muscle biopsy and definite findings on muscle MRI. Patients in the younger generation of family B and two members of families C (C:IV-4) and D (D:IV-8) did not have respiratory symptoms at ages 32–67 years but had evident findings on muscle biopsy and MRI. Cardiomyopathy was not manifest in any of the patients, based on clinical, electrocardiography, chest x-ray examinations or echocardiography (nine patients). Creatine kinase activity was normal or slightly elevated. EMG was myopathic with normal nerve conduction studies in 17 patients. EMG was reported as neurogenic in one patient (B:III-3) and with mixed findings in two (B:III-10, G:II-2).
There were 10 siblings in the second generation (figure 1, B:II) of family B, of whom six deceased patients had been affected. All four neurologically examined patients (B:II-3, -7, -9, -12) had documented muscle weakness and muscle biopsy obtained in two of them showed rimmed vacuolar myopathy and dystrophic findings. Three of the siblings (B:II-2, -4, -7) had respiratory failure, needed mechanical ventilation and died on average aged 65 years.
In Italian family J the proband with homozygous mutation had relatively early adult onset of respiratory failure and the heterozygous parents were reported healthy at ages 52 years and 55 years, which is compatible with a recessive mode of inheritance.7 After the same mutation in heterozygous state was identified in two siblings in family L of Portuguese origin causing the same disease but with considerably later onset, additional studies were performed by MRI in the parents of Italian family J (figure 2). Although subjectively healthy and with no muscle weakness on clinical testing, muscle MRI revealed clear pathology compatible with the known pattern of muscle involvement in HMERF. In French-Portuguese family K, the parents of the homozygous proband were first cousins with no signs of disease on clinical examination at the age of 57 years and 61 years, respectively. Muscle MRI was not performed. The parents of two siblings in family L were reported to be healthy but muscle MRI was not available.
The distribution and degree of fatty degenerative changes in muscles of 18 patients were evaluated (table 1). Semitendinosus muscle was moderately to severely affected in all patients and obturator, sartorius and gracilis muscles were similarly involved in most. Other frequently affected muscles were gluteus minimus and iliopsoas. Changes in other pelvic and thigh muscles were more variable and quadriceps and biceps femoris were relatively spared. In the lower legs gastrocnemius medialis and lateralis, as well as soleus muscles were relatively preserved, while in all other muscles the changes were moderate to severe. The homozygous proband of family J (J:II-1) had the typical phenotype with pathognomonic findings on imaging. His heterozygous parents without clinical symptoms had mild to moderate fatty degenerative changes on MRI; the father particularly in semitendinosus muscles and the mother in obturator muscles (figure 2).
Muscle histopathology and immunohistochemistry
Typical pathological findings were fibre-size variation and increase of internal nuclei (table 2). CBs were observed in all samples, although in some samples only in a few fibres. Rimmed vacuolar pathology was another constant feature but rimmed vacuoles did not appear in the same fibres with CBs. CBs were present in the sections of four out of the five samples available for additional immunohistochemical evaluation, ranging from one or a few fibres to 10–15% fibres harbouring CBs. They were generally found in subsarcolemmal position, often forming subsarcolemmal rings, and were different in size, but all displayed similar immunohistochemical features (figure 3). In particular, CBs were positive with antimyotilin, anti-αB-crystallin antibodies, and also contained actin and dystrophin. Desmin was absent from the core of the bodies but sometimes positive in a thin surrounding halo and detectable in areas of myofibrillar disarray. These myofibrillar disruption areas were also present in other fibres in the central position, and were positive with antimyotilin and anti-αB-crystallin antibodies. However, CBs were not reactive with antititin antibodies, and TDP-43 (TAR DNA binding protein-43) and p62/SQSTM1 present in rimmed vacuoles were absent from them as well. p62/SQSTM1 also showed a dotted appearance in some hypotrophic fibres, and was positive in the areas of myofibrillar disarray between CBs. CBs did not display affinity for ubiquitin, whereas ubiquitin positivity was sometimes detected at the periphery of the bodies and in abnormal fibres as a diffuse cytoplasmic increase. The autophagosome marker LC3 labelled rimmed vacuoles but was largely absent from CBs.
Haplotype segregation using markers at the TTN locus 2q31 was identified in all familial materials available. In the French family (Family A) all the affected patients shared the same haplotype that was not present in any of the healthy members of the family. The Finnish family (B) showed segregation of a different haplotype in the patients, which was also partially present in the Swedish (C and D), one British (E), one Italian (F) and one Argentinian (G) families. All patients in these families with the C30071R mutation shared a haplotype including markers D2S300 and D2S385. The shared haplotype was less than 1.3 Mb in size (less than 1.1 cM). Patients in one Italian family (J) and in two French families with Portuguese ancestry (K and L) showed segregation of yet another haplotype which included markers D2S300 and D2S385 and was less than 1.3 Mb in size (less than 1.1 cM). In Italian family J and French (Portuguese) family K both probands showed this identical short haplotype on both chromosomes with extended haplotype marker allele sharing in family K consistent with the parents being first cousins.
Exome sequencing was performed on two affected and one healthy member of French family A. When variants shared by both affected patients were filtered against dbSNP132 and the healthy member of the family, only one variant was found within the linked haplotype in the TTN gene. The variant g.274367C>G was located in TTN exon 343 and it caused one amino acid change, p.P30068R.
Since the new mutation in the French family and the previously reported HMERF A-band mutation3 ,4 were both in TTN exon 343, all other families were screened for exon 343 and four more mutations were identified (table 3). The causative mutation was identified in every patient with HMERF in our series. One of these mutations g.274375T>C (p.C30071R) previously reported in a few Swedish and UK families3 ,4 was now identified in one Finnish, one UK, one Italian, one Argentinian and in the two Swedish families. One of the new mutations g.274436C>T (p.P30091L) has been observed in an exome sequencing project in one single patient, but without further confirmation of its possible pathogenicity.14 This mutation was now identified in heterozygosity in two sibs in one French family, and in homozygosity in the proband of the second French family with Portuguese ancestry, as well as in homozygosity in the proband of Italian family J. The German mutation g.274426T>C (p.W30088R) and the new British mutation g.274428G>C (p.W30088C) were identified by direct sequencing of the candidate region and could be directly associated because of the identical phenotype and segregation in the affected patients only. Mutations were not present in dbSNP132, 1000 Genomes or NHLBI (the National Heart, Lung, and Blood Institute, Maryland, USA) Exome Sequencing Project databases. TTN exon 343 was sequenced from 102 Finnish and 96 Italian healthy controls and none of them had any of these mutations.
All five mutations are missense mutations changing one amino acid in the protein. The p.P30068R mutation changes a hydrophobic amino acid to a positively charged amino acid, the p.C30071R mutation changes a small neutral amino acid to a large positively charged amino acid, the p.W30088C mutation changes a large amino acid to a small amino acid which may affect cysteine-cysteine bindings, the p.W30088R mutation changes a hydrophobic amino acid to a positively charged amino acid, and the p.P30091L mutation changes a rigid amino acid to a flexible amino acid. All of the mutations are located in the same FN3 119 domain (A150) in A-band titin (figure 4).
HMERF disease was first described in single families more than 20 years ago.6 ,15 Two of these families showed linkage to the chromosome 2q31 locus16 and the first titin mutation associated with HMERF was identified in the M-band kinase domain.2 The genetic cause in the remaining families was elusive until new studies using exome sequencing eventually identified a new dominant titin mutation in three Swedish families and in three UK families in the distal part of A-band titin.3 ,4 Since both titin mutations showed an identical pattern of muscle involvement on MRI3 ,4 and histopathology, these parameters were used to reassess two previously reported families,6 ,7 two of the original Swedish families without established genetic diagnosis, and other unreported HMERF families regarding possible titin mutations.
In our larger collection of patients with HMERF in 12 unrelated families from different Caucasian populations one previously reported and four novel mutations were identified. Our findings emphasise the geographically wide occurrence, the importance of titin as a causative gene of HMERF disease and particularly the role of exon 343 as a mutational hotspot region. Furthermore, this study considerably expands our understanding of the clinical presentation and provides new insight into molecular pathology.
Previous linkage studies in French family A suggested that the titin gene locus was excluded.17 New muscle MRI studies lead to reclassification of some of the individuals in the family. Because of the identical pattern of muscle involvement with the other families with HMERF, A-band titin Sanger sequencing was also performed, besides exome sequencing, resulting in the novel p.P30068R mutation to be identified. As expected, after reclassification, new linkage studies showed that the disease was indeed linked to the titin locus.
The p.C30071R mutation, now identified in six families from five different populations, and previously reported in three Swedish and three British families,3 ,4 is the most frequent mutation in HMERF so far. The mutation mediates a dominant effect with full penetrance. The age of onset with this mutation varies from 16 years to 53 years being usually after age 30–35 years. All these families share a genomic region less than 1.3 Mb in size suggesting an ancestral founder. The p.P30068R, p.W30088C and p.W30088R mutations are novel and occur in single families with dominant inheritance and full penetrance. The age of onset and disease severity in patients with these new mutations is within the range of the phenotypes with the p.C30071R mutation. However, the fifth mutation identified, p.P30091L, in families J, K and L, shows a different penetrance and is neither completely dominant nor completely recessive. The parents of the probands in families J and K showed no signs of muscle disease and therefore the probands appeared to be sporadic patients. However, because the same mutation occurred in heterozygous state in family L, although with a milder phenotype, additional studies were needed. In families J and K the homozygous p.P30091L probands (J:II-1 and K:II-1) have a more severe disease with earlier onset and more rapid progression than the heterozygous patients of family L. Furthermore, in family J the heterozygous parents had definite signs of subclinical muscle disease on MRI with fatty degenerative change in HMERF typical muscles. In family L, the heterozygous patients had respiratory insufficiency although at much later age and without clear limb muscle weakness. Their muscle MRI findings were also milder. Since this novel p.P30091L mutation may or may not cause clinically manifest disease in heterozygous state, and it causes a clearly more severe phenotype in homozygosity, we prefer to call this mutation semirecessive or semidominant. The new mutations reported are currently undergoing further functional studies, although, segregation with the phenotype in reported families and the fact that after the first submission three other new mutations in the same exon were reported,18 ,19 make all these mutations very likely pathogenic.
The most typical feature of HMERF is respiratory failure at mid-adulthood in an ambulant patient at the first visit. In the case of absent or undiagnosed respiratory symptoms, more specific clues to enable correct diagnosis can be obtained from muscle imaging and histopathology. Our results on muscle imaging so far confirm the characteristics and pathognomonic pattern of muscle involvement in this disease.3–5 ,7
Muscle pathology is another key to diagnosis. CBs, a hallmark of the disease, were observed in all samples. However, in some biopsies they were present in one or a few fibres only and could be easily overlooked in the first reading. Rimmed vacuolar pathology was another consistent finding but these focal degenerative changes did not occur in the same fibres that contained CBs. Myofibrillar changes, Z-disk alterations and CBs that bind phalloidin (a marker for F-actin), and also contain desmin, myotilin, αB-crystallin, VCP (valocin-containing protein), and dystrophin, have been demonstrated in previous HMERF studies.1 ,3 ,4 ,7 ,17 ,20 Moreover, p62/SQSTM1 containing cytoplasmic inclusions has been observed.2 In our new series of immunohistochemistry studies including three different mutations (two dominant and the semirecessive) we were able to detail the different types of accumulations and aggregations in three different cytoplasmic abnormalities: (1) CBs contained dystrophin, myotilin, actin and αB-crystallin; (2) rimmed vacuoles contained ubiquitin, TDP-43, p62/SQSTM1 and LC3 positive components; and (3) in regions with myofibrillar disarray between CBs or centrally in some fibres the expression of desmin, ubiquitin and p62/SQSTM1 was increased. Surprisingly, the most compelling finding was that CBs were not reactive for titin. Some features in patients with HMERF overlap with myofibrillar myopathies (MFM). However, the dark and hyaline changes in the cytoplasm of affected fibres on trichrome staining characteristics for MFM are largely lacking in HMERF. Instead the routine stainings are showing CBs which are not a hallmark of MFM.21
The spectrum of different human titinopathies is growing but none of the other described forms8–10 ,22 show the unique pathology of fibres with multiple CBs apart from rimmed vacuolated fibres, which are both caused by the mutant titin protein. Immunohistochemistry findings indicate that the mutant protein itself is not aggregating and not seeding the CBs but seems to trigger the aggregation of other sarcomeric proteins. The first titin mutation causing HMERF in the kinase domain of M-line titin leads to disruption of the kinase associated protein complex with Nbr1, p62/SQSTM1 and MuRF2.2 This was shown to lead to the mislocalisation of the multiprotein complex, which is involved in ubiquitin-mediated regulation of transcription, protein turnover via the ubiquitin proteasome system and autophagy-mediated protein turnover via the interaction with ubiquitinated proteins and LC3 of p62/SQSTM1 and nbr1.23 However, this kinase mutation has later been shown to occur as a variant (rs140319117) with a frequency of 0.0018 among European Americans that warrants comprehensive new assessment of the mechanism. The A-band region of titin has a central role in controlling myosin thick filament positioning and function,24 ,25 but the specific links of this particular FN3 119 domain (A150) to dysregulated protein recycling and autophagy turnover remain to be clarified.
Five different mutations causing HMERF disease in various populations identified in exon 343 of TTN makes this a first-step target for molecular diagnosis and the range of mutations indicates that HMERF disease is not extremely rare and is not restricted to the Scandinavian population.
Dr R Carlier and Dr N Romero, Myology Institute collected data and provided general advice. Pr D Hénin, Claude-Bernard Bichat Hospital, Department of Pathology, Genethon and AFM collected data and provided general advice. Dr E Bertini, Department of Neurosciences, Unit of Molecular Medicine, Bambino Gesu Children's Hospital collected data and provided general advice. Dr E Ricci Institute of Neurology, Catholic University School of Medicine, Rome, Italy collected data and provided general advice. Dr Maria Saccoliti MD, Department of Neuropathology, Institute for Neurological Diseases, FLENI, Buenos Aires, Argentina performed electron microscopy. MSc Maritta Kariniemi, Neuromuscular Research Unit, Tampere Finland assisted with the preparation of figure 3.
JP and AE contributed equally.
Contributors Study conception and design: AO and BU. Acquisition of data: JP, AE, FC, GT, BB, BE, AE-L, JL, MK, IM, RQ, MD, AB, ALT, JAB, MT, PG, FC, FN, RS, LE, AO, PH and BU. Analysis and interpretation of data: JP, AE, FC, GT, FX, BE, PL, JT, TR, CS, CH, RS, LE, AO, PH and BU. Drafting of the manuscript: JP, AE, PH, BU. Critical revision of the manuscript for important intellectual content: FC, GT, FX, BB, BE, AE-L, JL, MK, IM, RQ, PL, MD, AB, ALT, JAB, MT, PG, FC, CS, FN, CH, RS, LE, AO, PH and BU. Obtained funding: BU. Administrative, technical and material support: PH and BU. Study supervision: PH and BU.
Funding This study was supported by the Folkhälsan Research Foundation, the Academy of Finland, the Sigrid Jusélius Foundation, the Liv o Hälsa Foundation and Tampere University Hospital Research Funds (BU).
Competing interests RQ has received consultancy fee from Global, lecture fee from Genzyme, and Muscular Dystrophy Campaign grant. PL has received grants and honorarium from Genzyme Company. He is a member of the Pompe advisory Board for Genzyme. JT and TR have received funding from Academy of Finland, and FX has received funding from ALF-grant (medical training and clinical research). JL's institution has received grants from ANR and AFM. He is employed by INSERM. All other authors report no conflicts of interest.
Patient consent Obtained.
Ethics approval Systemic collection of clinical data and all genetic studies in Finland were approved by the Ethics committee of Tampere University Hospital, Finland.
Provenance and peer review Not commissioned; externally peer reviewed.