Background
Hereditary myopathy with early respiratory failure (HMERF) also known as Edström myopathy is a disorder manifesting with predominantly proximal muscle weakness of the lower and upper extremities with respiratory insufficiency and involvement of neck flexors early in the disease course [
1]. A number of cases showing autosomal dominant pattern of inheritance and sporadic cases have been reported [
2‐
4]. A distinctive pathological feature of HMERF is the presence of cytoplasmic bodies. The first
TTN mutation causing HMERF has been identified by Lange and colleagues [
5] in Swedish families originally described by Edström et al. [
1]. This mutation was defined as p.Arg279Trp by using residue numbering according to the crystal structure of the titin kinase domain [
6]. Based on updated databases (GenBank NP_001243779 and UniProt Q8WZ42), the mutant residue has been re-numbered as p.Arg32450Trp [
7]. Recently, HMERF in several newer North European families have been associated with a g.274375T>C: p.Cys30071Arg mutation in the A-band of titin [
7,
8].
TTN mutations have been known to cause other neuromuscular and cardiac disorders, among them dilated cardiomyopathy type 1G [
9] and neuromuscular disorders such as tibial muscular dystrophy (TMD), or Udd myopathy [
10], limb–girdle muscular dystrophy type 2J (LGMD2J) [
11,
12], and autosomal recessive early-onset myopathy with fatal cardiomyopathy (EOMFC) [
13]. Mutations associated with dilated cardiomyopathy are overrepresented in the titin A-band [
14]; the mutation identified in Swedish HMERF families by Lange et al. [
5] is in the titin protein kinase domain, while mutations for TMD, LGMD2J, and EOMFC are located in the C-terminal end of the M-band.
Titin is the largest muscle protein known, a filamentous molecule stretching for half-sarcomere, from the Z-disk (N-terminus) spanning the A-band and extending to the M-band (C-terminus) [
15]. Titin isoform of skeletal muscle is composed of >33,000 amino acids, weighs 3,700 kD, and its length is 2 μm [
16]. Titin has a modular structure; up to 90% of its mass consists of repeating immunoglobulin-like (Ig) and fibronectin type III (FN3) domains. Titin serves as a molecular template for the assembly of the myosin-based filament and is responsible for the stabilization of the thick filament and the structural integrity of the entire sarcomere by acting as a scaffold [
17]. Titin is a molecular spring that provides elasticity to the sarcomere and ensures its return to the original length after muscle relaxation [
18]. Titin is also involved in signal transduction from the myofibrils to other compartments of the muscle cell, including the nucleus [
19].
The elastic part of titin is its I-band region composed of 40 Ig-domains; the A-band is stable and not extensible due to its strong interaction with the thick filament [
20,
21]. The A-band is composed of stretches of FN3 domains interspaced by single Ig domains, forming the unique titin super-repeat architecture. The N-terminal super-repeat within the A-band (D-zone) comprises six copies of a 7-element structure arranged as Ig-(FN3)2-Ig-(FN3)3. The second super-repeat located C-terminally (C-zone) is organized into eleven copies of an 11-element motif arranged as Ig-(FN3)2-Ig-(FN3)3-Ig-(FN3)3 [
22]. The two super-repeats of the A-band region provide regularly spaced binding sites for components of the thick filaments [
17,
23,
24] and serve as a molecular ruler that regulates the assembly and the length of the thick filament [
25]. FN3 elements provide binding sites for myosin, while Ig-like domains may be responsible for interaction with other ligands. A-band is evolutionarily conserved, unlike the Z-disk and I-band segments of titin that are highly divergent [
18]. Titin kinase domain contains a catalytic domain and an auto-regulatory C-terminal tail [
6], which wraps the active site of the catalytic domain. The p.Arg32450Trp mutation is located at the N-terminal helix (alphaR1) of the kinase domain [
5].
TTN gene is positioned in the 2q31 chromosome region and consists of 363 exons.
In the process of genetic testing of patients with myofibrillar myopathy (MFM) we encountered familial and sporadic cases of skeletal myopathy with or without associated respiratory abnormalities who lack mutations in MFM-associated genes [
26,
27]. To identify a causative mutation in affected members of a large U.S. family suffering from proximal myopathy and respiratory failure we carried out whole exome sequencing and determined that the mutation was in the
TTN gene. Screening of other families led to the identification of a similar mutation in affected individuals from two other families originating from a Native American population in Canada and from Spain, indicating that missense mutations in
TTN are the cause of HMERF in families of divergent origins. We compared phenotypic features of HMERF in three families under our study with previously reported clinical/pathological descriptions of the disease caused by three
TTN mutations in various populations and refined diagnostic criteria of HMERF. Both p.Gly30150Asp and p.Cys30071Arg
TTN mutations disrupt a fibronectin type III element of titin A-band super-repeat designated as A150.
Discussion
After Edström et al. [
1] outlined a new disorder based on analysis of patients with an adult-onset proximal myopathy and early respiratory muscle weakness, a number of similar cases were described. The clinical/pathological definitions of this disease later named HMERF were significantly refined and extended [
2,
3]. Two Swedish families originally reported by Edström et al. were genotyped, the disease-associated locus assigned to chromosome 2q31 in the vicinity of the
TTN gene [
34], and eventually a p.Arg32450Trp mutation in the kinase domain of titin was discovered [
5]. Further attempts at finding HMERF-causing
TTN mutations were hampered by the enormous size of this gene. Next generation sequencing opened up opportunities for new discoveries. Using this technology, the p.Cys30071Arg
TTN mutation was recently identified as the cause of HMERF in North European families [
7,
8].
Whole exome sequencing performed on five affected members of Family A showed two variants, c.175C>T:p.Arg59Cys in
CDC2 on chromosome 10q21.1 and c.90674G>A:p.Gly30150Asp in
TTN on chromosome 2q31 as having the highest deleteriousness scores. Both are segregating with the disease.
CDC2 variant was excluded on the basis that it was not evidently deficient in western analysis and had been recorded in the general population thus suggesting that this variant is rather a rare benign polymorphism (
rs8755). The
TTN p.Gly30150Asp change is located in the highly conserved titin A-band and not present in any available SNP database. The final selection of the
TTN mutation was helped by reports of finding the HMERF-associated p.Cys30071Arg mutation in North European families [
7,
8]. Screening of 45 additional familial and sporadic patients in which the role of MFM-associated genes have been excluded led to the identification of the
TTN p.Cys30071Arg mutation in the index patient of family B originating from Native American population; this same mutation was found in a HMERF patient from Spain, indicating that missense mutations in
TTN are the cause of HMERF in diverse populations.
We compared the disease phenotypes in groups of HMERF patients originating from different populations and carrying various
TTN mutations (Table
4). Although the number of patients in each column is small for a comprehensive comparison, there is enough data to summarise the basic features of this disease. The pattern of inheritance is autosomal dominant. The age of disease onset is in the teens or the 20-s in most patients; it was at somewhat older age in the U.K. family [
2,
7]. Pelvic and shoulder girdle weakness was the earliest and predominant sign in every patient, with several exceptions in the family from the U.K. which showed weakness in distal muscles at presentation. In the U.S. family, we observed characteristic sternocleidomastoid/trapezius muscle weakness and atrophy without scapular winging, and calf hypertrophy that was evident before muscle wasting has developed with disease progression. Respiratory function was affected early in the disease: about a third of patients presented with respiratory insufficiency at the disease onset or developed during the first year of illness, and two thirds had moderate respiratory failure requiring ventilation support before the fifth year of illness while they were still ambulant. With disease progression, most of the patients developed weakness in the ankle dorsiflexors. None of the HMERF patients had cardiac involvement, although mutations in
TTN are known to cause dilated and hypertrophic cardiomyopathy [
14]. This may depend on differential expression of tissue-specific titin isoforms. Creatine kinase (CK) levels were normal or slightly elevated. Muscle imaging in our patient showed characteristic and diagnostically significant abnormality in iliopsoas, obturator externus, semitendinosus, gracilis and sartorius muscles on the thigh level and the peroneal group at the mid-lower-leg level, which is in agreement with previous observations [
3,
8]. Disease outcome for patients with HMERF is one of the worst known for an autosomal dominant adult-onset myopathy because it leads to incapacity, wheelchair dependency, and the need for permanent ventilatory support relatively early in life.
Table 4
Comparative clinical characteristics of reported patients with confirmed titin mutations
Inheritance pattern | AD | AD | AD | AD | AD |
Country | Sweden | Sweden | Britain | U.S. | Spain/Canada |
No. of studied patients | 7 | 8 | 22 | 5 | 2 |
Onset age, mean (range), yrs | 24.8 | 28.8 | 44.6 | 19 | 29 |
(14–40) | (18–40) | (22–71) | (13–29) | (22–36) |
Gender (woman/man) | 5/2 | 5/3 | 12/10 | (0/5) | 2/0 |
Distribution of muscle weakness
| | | | | |
Pelvic girdle | 7/7 | 8/8 | 15/21 | 5/5 | 2/2 |
Shoulder girdle | 7/7 | 4/7 | 10/22 | 3/5 | 1/2 |
Neck flexors | 7/7 | 8/8 | 7/22 | 5/5 | 2/2 |
Trunk muscles | Nr | 8/8 | Nr | 3/5 | 1/2 |
Knee flexors | Nr | 7/8 | 7/22 | 3/5 | 1/2 |
Knee extensors | Nr | 2/8 | 6/22 | 3/5 | 2/2 |
Ankle dorsiflexors | 2/7 | 7/8 | 17/22 | 5/5 | 2/2 |
Plantar flexors | Nr | 1/8 | 5/22 | 3/5 | 0/2 |
Finger extensors | 4/7 | 1/8 | 4/22 | 3/5 | 2/2 |
Finger flexors | Nr | 1/8 | 3/22 | 2/5 | 0/2 |
Respiratory failure | 7/7 | 8/8 | 12/22 | 2/5 | 2/2 |
Mildly elevated CK | 1/7 | 1/1 | 11/19 | 1/1 | 2/2 |
Myopathic EMG | 7/7 | 1/1 | 11/18 | 1/1 | 2/2 |
Outcome
| | | | | |
Wheelchair dependency | Nr | 1/8 | 2/22 | 2/5 | 1/2 |
Ventilation dependency | 4/7 | 4/8 | 6/22 | 2/5 | 2/2 |
Death before age 65 | Nr | 1/1 | Nr | 2/2 | 1/1 |
The most characteristic abnormalities on muscle biopsies of patients with HMERF are the small round well defined eosinophilic cytoplasmic bodies. Patients with advanced disease show in addition large diffuse polymorphic accumulations of material corresponding to protein aggregates that contain F-actin, dystrophin, myotilin, and filamin C as we determined by immunohistochemistry and as shown in other series of HMERF patients. In agreement with previous observations [
4], desmin was focally increased under the sarcolemma and at the periphery of cytoplasmic bodies but not within them. Interestingly, we detected TDP-43 aggregates in the sarcoplasm of several fibers. TDP-43 (TAR-DNA-binding protein-43) is typically found in cytoplasmic inclusions of patients with motor neuron disease and frontotemporal lobar degeneration [
35]. TDP-43-stained aggregates were also found in affected muscle fibres in rimmed vacuolar myopathies, including inclusion body myositis, GNE myopathy, myofibrillar myopathies, and oculopharyngeal muscular dystrophy [
36‐
38].
Ultrastructural analysis in two of our patients revealed early disintegration of the Z-disks followed by myofibrillar dissolution and aggregation of degraded filaments. Taken together, the immunohistochemical and ultrastructural features are closely reminiscent of those seen in MFM and based on these morphological characteristics HMERF can be considered a member of the same large group of muscle diseases recently named protein aggregate myopathies (PAMs) [
39]. While differential diagnosis of HMERF is solidly based on the presence of cytoplasmic bodies, it needs to be indicated that cytoplasmic bodies observed in HMERF-affected muscles are reminiscent of and should be distinguished from the globular cytoplasmic inclusions seen in muscle biopsies of a fraction of patients with Pompe disease [
40]. This is of major importance since early respiratory weakness is also frequently observed in patients suffering from the late-onset Pompe disease [
41]. Strong acid phosphatase activity and lack of filaments or any other Z-disc components in Pompe disease structures [
40] may help to distinguish them from cytoplasmic bodies found in HMERF. Accumulation of cytoplasmic and occasionally intranuclear tubulofilaments have been noted in patients with MFM, particularly myotilinopathy [
42‐
44] and C-filaminopathy [
45,
46], demonstrating that the boundaries between HMERF and MFM need to be defined more clearly. Indeed, the combination of proximal and distal weakness and respiratory failure, muscle imaging results and myofibrillar inclusions on muscle histopathology make HMERF similar to desminopathy or alphaB-crystallinopathy. A major distinguishing feature is the lack of cardiac involvement which is seen in the majority of patients with desminopathy or alphaB-crystallinopathy [
47] and, importantly, granulofilamentous material which constitutes the ultrastructural hallmark of desminopathy and alphaB-crystallinopathy [
27,
44,
48] is not a feature of HMERF.
There is very limited data on possible consequences of disruptions in FN3 elements within the conserved titin A-band region, although there has been a report of a mutation occurring within FN3 domains in patients with arrhythmogenic right ventricular cardiomyopathy [
49]. A three-dimensional structure of an FN3 domain revealed that many of its conserved residues are exposed on the surface of the domain; most likely, they serve as binding sites specifically binding sub-fragment 1 of myosin [
20]. Residues replaced by HMERF mutations are known to be highly conserved, specifically the cysteine residues that are the most conserved throughout the entire A-band structure [
20]. If the previously proposed structural model of A77 [
33] reflects the structure of other A-band FN3 domains, the mutated residues Cys30071 and Gly30150 are located in the side chain that most likely serve as a binding site. Some indications regarding the pathomechanisms in HMERF come from the results of experimental manipulations with FN3 domains that led to an increased rate of molecular aggregation [
50].
The p.Arg32450Trp mutation in the kinase domain identified in Swedish families may affect its ability to send signals regarding overextention and other abnormal physiological events in the contracting muscle [
51]. Specifically, the mutation disrupts the interaction of titin with NBR1 protein, a ubiquitin-binding scaffold protein, and in turn with p62/SQSTM1 and MURF2. The NRB1/P62 complex plays an important role in autophagic and proteasome degradation of ubiquitinated proteins [
5]. The affected muscle shows disrupted myofibrils, NBR1 localization in cytoplasmic bodies, and aggregation of the p62/SQSTM1 complex with myofibrillar proteins. This suggests that the p.Arg32450Trp mutation affects the ability of titin to control muscle protein turnover and through this action disrupts muscle sarcomere maintenance.
Competing interests
The authors declare no competing interests.
Authors’ contributions
CT contributed to writing the manuscript, the study concept and design, the analysis and interpretation of data. MO contributed to the study concept and design, clinical and pathological studies, writing the manuscript and the analysis and interpretation of data. MCD contributed to the study concept and design, clinical and pathological studies, writing the manuscript and the analysis and interpretation of data. KS, JMB, FT, NV, EF contributed to the study concept and design, generating of data, clinical evaluations, pathological studies, and the analysis and interpretation of data. NS contributed to the study concept and design, carried out genetic studies, participated in analysis and interpretation of data. LGG conceived the study, contributed to the study concept and design, writing the manuscript and the analysis and interpretation of data. All authors approved the final version of the manuscript.