OUP user menu

Novel deletion of lysine 7 expands the clinical, histopathological and genetic spectrum of TPM2-related myopathies

Ann E. Davidson, Fazeel M. Siddiqui, Michael A. Lopez, Peter Lunt, Heather A. Carlson, Brian E. Moore, Seth Love, Donald E. Born, Helen Roper, Anirban Majumdar, Suman Jayadev, Hunter R. Underhill, Corrine O. Smith, Maja von der Hagen, Angela Hubner, Philip Jardine, Andria Merrison, Elizabeth Curtis, Thomas Cullup, Heinz Jungbluth, Mary O. Cox, Thomas L. Winder, Hossam Abdel Salam, Jun Z. Li, Steven A. Moore, James J. Dowling
DOI: http://dx.doi.org/10.1093/brain/aws344 508-521 First published online: 14 February 2013

Summary

The β-tropomyosin gene encodes a component of the sarcomeric thin filament. Rod-shaped dimers of tropomyosin regulate actin-myosin interactions and β-tropomyosin mutations have been associated with nemaline myopathy, cap myopathy, Escobar syndrome and distal arthrogryposis types 1A and 2B. In this study, we expand the allelic spectrum of β-tropomyosin-related myopathies through the identification of a novel β-tropomyosin mutation in two clinical contexts not previously associated with β-tropomyosin. The first clinical phenotype is core-rod myopathy, with a β-tropomyosin mutation uncovered by whole exome sequencing in a family with autosomal dominant distal myopathy and muscle biopsy features of both minicores and nemaline rods. The second phenotype, observed in four unrelated families, is autosomal dominant trismus-pseudocamptodactyly syndrome (distal arthrogryposis type 7; previously associated exclusively with myosin heavy chain 8 mutations). In all four families, the mutation identified was a novel 3-bp in-frame deletion (c.20_22del) that results in deletion of a conserved lysine at the seventh amino acid position (p.K7del). This is the first mutation identified in the extreme N-terminus of β-tropomyosin. To understand the potential pathogenic mechanism(s) underlying this mutation, we performed both computational analysis and in vivo modelling. Our theoretical model predicts that the mutation disrupts the N-terminus of the α-helices of dimeric β-tropomyosin, a change predicted to alter protein–protein binding between β-tropomyosin and other molecules and to disturb head-to-tail polymerization of β-tropomyosin dimers. To create an in vivo model, we expressed wild-type or p.K7del β-tropomyosin in the developing zebrafish. p.K7del β-tropomyosin fails to localize properly within the thin filament compartment and its expression alters sarcomere length, suggesting that the mutation interferes with head-to-tail β-tropomyosin polymerization and with overall sarcomeric structure. We describe a novel β-tropomyosin mutation, two clinical-histopathological phenotypes not previously associated with β-tropomyosin and pathogenic data from the first animal model of β-tropomyosin-related myopathies.

  • nemaline
  • myopathies
  • muscle and nerve pathology
  • mutation
  • neuromuscular disorders

Introduction

β-Tropomyosin (β-Tm or TPM2) is an actin-binding protein that functions to stabilize filamentous actin and to regulate calcium-dependent actin-myosin interactions (Kee and Hardeman, 2008). It is one of four tropomyosin isoforms, is preferentially expressed in slow (type I) skeletal muscle and is the major tropomyosin isoform expressed during skeletal muscle development (Laurent-Winter et al., 1991). TPM2 is a key component of calcium-mediated muscle contraction, serving physically to block actin-myosin interaction in the relaxed state (Engel and Franzini-Armstrong, 2004). TPM2 exists preferentially in heterodimers with either TPM1 or TPM3, though it can also form homodimers (Perry, 2001). Tropomyosin dimers containing TPM2 additionally undergo head-to-tail polymerization, ultimately spanning the length of the thin filament in close association with actin (Perry, 2001; Murakami et al., 2008).

Heterozygous mutations in TPM2 have been described in nemaline myopathy (Donner et al., 2002; Tajsharghi et al., 2007b), cap myopathy (Lehtokari et al., 2007; Ohlsson et al., 2008; Clarke et al., 2009) and in distal arthrogryposis types 1A and 2B (Sung et al., 2003; Tajsharghi et al., 2007a), with cap myopathy being the most commonly reported TPM2-related muscle disease. In total, nine dominant mutations and one recessive mutation have thus far been identified, with the single recessive case a homozygous null allele associated with Escobar syndrome (a non-lethal form of multiple pterygium syndrome) (Monnier et al., 2009). There is significant clinical heterogeneity between affected individuals with TPM2 mutations, and no consistent clinical picture has emerged to unite the reported cases. In terms of pathophysiology, known myopathy-associated mutations in TPM2 are predicted to change the affinity of β-Tm dimers for F-actin or impair responsiveness to calcium (Ochala, 2008; Ochala et al., 2008; Marttila et al., 2012), whereas the single distal arthrogryposis-associated mutation studied to date was demonstrated to alter actin-myosin cross-bridge dynamics in isolated patient myofibres (Ochala et al., 2007, 2010). There is also evidence to suggest that some TPM2 mutations may alter the relative composition of tropomyosins in the skeletal muscle thin filament (Nilsson and Tajsharghi, 2008).

Nemaline myopathy and cap disease are part of a larger group of disorders referred to as congenital myopathies (Nance et al., 2012). This is a heterogeneous group of primarily childhood onset diseases typically characterized clinically by congenital hypotonia and weakness; each separate disorder is largely defined by the different structural abnormalities observed on muscle biopsy (Jones et al., 2003). The most common subtypes are nemaline myopathy, core myopathy and centronuclear myopathy (Amburgey et al., 2011). Core-rod myopathy, like cap disease, is a rare myopathy variant and is defined by muscle biopsies containing both nemaline rods and cores (Nance et al., 2012). To date, mutations in RYR1 (von der Hagen et al., 2008; Hernandez-Lain et al., 2011), NEB (Romero et al., 2009), ACTA1 (Kaindl et al., 2004), CFL2 (Agrawal et al., 2007) and KBTBD13 (Sambuughin et al., 2010) have been identified in association with core-rod myopathy; it is clear, however, that additional genetic causes remain to be determined for many individuals.

Distal arthrogryposis syndromes are defined by the presence of contractures detected at birth in more than one body area and in a distal predominant pattern, and are not classically considered congenital myopathies (Bamshad et al., 2009; Kimber et al., 2012). Many individuals with distal arthrogryposis, however, have true muscle weakness and may have biopsies with abnormalities similar to those in congenital myopathies (Sewry et al., 2008). Thus, the boundaries between the two entities are becoming increasingly blurred. Distal arthrogryposis syndromes are delineated into 10 subgroups based on the distribution and severity of contractures (Bamshad et al., 2009). Heterozygous TPM2 mutations are associated with distal arthrogryposis type 1A (characterized primarily by camptodactyly and clubfoot) (Sung et al., 2003) and distal arthrogryposis type 2B (Sheldon-Hall syndrome, characterized by triangular face, down-slanting palpebral fissures, a small mouth and mandible, cervical webbing, camptodactyly and clubfoot) (Tajsharghi et al., 2007a). Distal arthrogryposis type 7, up to now uniquely associated with heterozygous mutation in the MYH8 gene, presents clinically as trismus-pseudocamptodactyly (Toydemir et al., 2006), a combination of clinical features including limited mouth opening (trismus) and contracture of the long flexor tendons of the fingers revealed clinically on full extension at the wrist (pseudocamptodactyly); in that position, patients are unable to hold the palms of the hands and fingers flat together (praying hand sign) (Hecht and Beals, 1969).

In this study, the goal at the outset was to establish the genetic basis of disease in a family with autosomal dominant core-rod myopathy and in three families with autosomal dominant distal arthrogryposis type 7 (trismus-pseudocamptodactyly syndrome) (Mabry et al., 1974). Surprisingly, affected individuals in all four families share a unique three base pair deletion (c.20_22del) in TPM2. This dominant mutation results in deletion of a highly conserved amino acid (p.K7del). The molecular consequences of this mutation were explored using simulated modelling and by expression studies in the zebrafish, in which the mutation appears to disrupt intermolecular interactions and alter mutant TPM2 distribution. Our findings expand the clinical, pathological and pathomechanistic spectrum of TPM2-related disorders.

Materials and methods

Approvals

Patient DNA was collected after obtaining informed consent using an institutional review board approved protocol. Experimentation in the zebrafish was performed under an Institutional Animal Care and Use Committee-approved protocol.

Clinical evaluation

Medical history was obtained, and physical examination was performed as part of routine neuromuscular clinical evaluations.

Muscle biopsies

Biopsies were obtained as part of the clinical diagnostic process. They were evaluated by standard light and electron microscopic protocols.

Western blot analysis

Muscle biopsies from the probands of Families 1 and 4 were evaluated by western blot analysis using published methodology (Anderson and Davison, 1999). Briefly, homogenates from pooled cryosections were run on both 3–15% gradient and straight 15% polyacrylamide gels. Tropomyosin isoforms were detected on a LiCor Odyssey® Infrared Imaging System using the mouse monoclonal anti-tropomyosin antibody, clone TM311 (Sigma-Aldrich) and the goat, anti-mouse secondary antibody, IgG IR 800 (LiCor).

Targeted gene testing

Candidate gene analysis was performed in clinical diagnostic laboratories using conventional Sanger sequencing methodology.

Targeted linkage analysis

For Family 1, linkage studies were performed by haplotype analysis of the following gene loci: ACTA1, KBTBD13, MYH7, NEB, RYR1 and TPM3.

Whole exome capture and next-generation sequencing

Whole exome capture was performed using the Nimblegen SeqCap EZ platform. Massively parallel sequencing was performed using an Illumina HiSeq2000 instrument. Sequence data were aligned to the reference genome using Burrows-Wheeler Aligner, and variants detected and called using previously described methodology. Mean coverage for each individual was >50×. The c.20_22del mutation was confirmed using Sanger dideoxy sequencing.

Computational modelling

PyMol (Version 1.5.0.4, Schrödinger, LLC) was used to computationally generate models of the p.K7del mutant of TPM2, based on the 3U59 crystal structure (Rao et al., 2012). The tail helices of the head-to-tail-packed polymer were created through propagating the unit cell of the crystal structure. The models of the p.K7del mutants were created using PyMol’s single-site mutagenesis tool to change residues 2–5, and residue 1 was deleted manually. The changes in protein–protein contacts were assessed by visualizing the protein surfaces. The changes were summarized schematically using wheel diagrams.

Zebrafish expression studies

Zebrafish TPM2 complementary DNA was obtained from Open Biosystems. It was subcloned in-frame with enhanced green fluorescent protein (GFP) by PCR into the Tol2/gateway expression system. Complementary DNA expression was driven by the 5′ CMV/SP6 promoter. Mutagenesis to engineer the p.K7del mutation was accomplished by site-directed mutagenesis according to the manufacturer’s instructions (Stratagene). Complementary DNA was injected into 1-cell stage zebrafish embryos. Myofibres were isolated and analysed from 4- and 7-day post-fertilization zebrafish. Immunostaining of myofibres was performed as previously described (Dowling et al., 2009, 2012; Telfer et al., 2012). Antibodies included α-actinin (1:100; Sigma) and α-tropomodulin-1 (1:100; ProteinGroup).

Results

Family 1

Clinical assessment reveals a distal myopathy with autosomal dominant inheritance

The proband from Family 1 was a 9-year-old male subject who presented with complaints of gait unsteadiness that began in early childhood. Physical examination was notable for bilateral calf and anterior tibialis atrophy, weakness most prominently of plantar flexion, flexed toes while walking and an inability to toe walk. By historical report and by direct examination of some family members, four additional individuals, all male subjects, across three generations, were also affected with similar symptoms of mild distal lower extremity atrophy and weakness (pedigree presented in Supplementary Fig. 1). Additional clinical features of affected members of the pedigree included wrist and Achilles tendon contractures, pronounced difficulty with mouth opening (i.e. trismus) and progressive gait disturbance in the oldest affected individual (who eventually lost the ability to ambulate). Three members of the family also have pseudocamplodactyly [abnormally short muscle-tendon units in the fingers, causing the fingers to curve or bend (camptodactyly) when the hand is dorsiflexed]. No cardiac or pulmonary problems were reported in any affected member of the family. A summary of clinical features is presented in Supplementary Table 1.

Muscle biopsy is consistent with a diagnosis of core-rod myopathy

A diagnostic quadriceps muscle biopsy was performed on the proband at age 9 years. It showed increased variation in fibre size with both atrophic and hypertrophic fibres. A small number of muscle fibres contained numerous internalized nuclei (Fig. 1A, B, and E). A type I fibre predominance was present (Fig. 1C1 and C2). Enhanced granular staining on Gomori trichrome (Fig. 1E) was shown to be large nemaline rods when the biopsy was evaluated by electron microscopy (Fig. 1G and I). Rare cap-like subsarcolemmal clusters of nemaline rods (nemaline bodies) were observed only on ultrastructural evaluation (data not shown). Central regions of irregular staining by haematoxylin and eosin and Gomori trichrome suggested the presence of cores (Fig 1A, B, and E). Oxidative stains revealed more numerous regions suggestive of cores (Fig. 1D and F). Core-like disruption of the sarcomeric cross striations was also visualized by desmin immunostaining (Fig. 1H). Ultrastructural evaluation confirmed the presence of numerous cores of variable size (Fig. 1J).

Figure 1

Quadriceps muscle biopsy of Proband 1 is consistent with core-rod myopathy. There is increased variation in fibre size, and a small number of muscle fibres contained numerous internal nuclei (A, haematoxylin and eosin). A type I fibre predominance is present (C1, slow myosin heavy chain; C2, fast myosin heavy chain). Enhanced granular staining on Gomori trichrome (E) is shown by electron microscopy to be large nemaline rods (G and I). Central regions of irregular staining by haematoxylin and eosin suggest the presence of minicores (B). Oxidative stains reveal more numerous regions suggestive of minicores (D, NADH (nicotinamide adenine dinucleotide) ; F, cytochrome c oxidase). Core-like disruption of the sacromeric cross striations is also noted by desmin immunostaining (H). Ultrastructural evaluation confirms the presence of numerous minicores of variable size (J). Scale bars: A, E, F and H = 100 µm; B = 50 µm; C1 and C2 = 300 µm; D = 200 µm; G and J = 6 µm; I = 1 µm.

Haplotype analysis excludes the known causes of core-rod myopathy

The known genetic causes of core-rod myopathy are mutations in the following genes: ACTA1, KBTBD13, NEB and RYR1. Mutations in these genes, as well as in TPM3 and MYH7, were excluded by targeted linkage evaluation using haplotype analysis (Supplementary Fig. 2). Direct sequencing of ACTA1 was also performed and did not reveal a causative mutation. Retrospective analysis of the TPM2 locus shows a common haplotype on one allele of each affected family member.

Whole exome capture and next-generation sequencing identifies a novel mutation in TPM2 as the cause of disease

To determine the genetic basis of disease in this family, whole exome capture followed by massively parallel sequencing was performed on genomic DNA from the proband and an affected uncle. Candidate variants were filtered based on the following assumptions: (i) they are shared by both affected family members; (ii) likely to be deleterious missense, nonsense or splice site mutations as predicted by PolyPhen2, SIFT or SeattleSNP; and (iii) not present in dbSNP, 1000 genomes, the NHLBI variant server, or our in-house exome variant database. A total of 266 variants met these criteria (186 single nucleotide variants, of which seven were nonsense and 18 predicted by PolyPhen2 to be probably damaging, and 80 insertion/deletions). Visual inspection of the list identified a three base pair deletion (c.20_22del) within the coding sequence of TPM2 that is predicted to cause in-frame deletion of the lysine at position 7 (p.K7del, Fig. 2). Conventional Sanger sequencing was performed to confirm that the mutation segregated with disease in the family (Fig. 2), and independent validation of the mutation in the proband was achieved in a clinical diagnostic laboratory. Of note, this lysine is invariantly conserved in all known TPM2 homologues and is also found to be conserved in TPM1 and TPM3 (Fig. 2). Only TPM4, which is not a known binding partner of TPM2 (Perry, 2001) and which likely performs functions distinct from TPM2 (Lin et al., 2008), does not have a lysine at this position.

Figure 2

Deletion of highly conserved lysine 7 in TPM2 in a family with core rod myopathy and two families with distal arthrogryposis type 7. (A) Schematic of the known TPM2 mutations along with their associated muscle conditions. The new mutation (LYS7DEL) is depicted in red. (B) Graphic depiction of amino acid (top) and genomic DNA sequence homology for TPM2 homologues as well as for TPM1, 3 and 4. (C) Sequence chromatogram of the proband from Family 1 (bottom) and his unaffected mother. Red bar highlights the start of the mutation.

Family 2

Clinical assessment identifies a family with autosomal dominant distal arthrogryposis type 7

The proband from Family 2 was a 15-year-old female subject who presented at birth with distal arthrogryposis. She had bilateral calcaneovalgus feet, right vertical talus, tight knee flexion and congenital hip dislocations. She also had ulnar deviation of her fingers with adduction of the thumb but no other contractures. She had normal interphalangeal flexion creases. At 5 years of age, she was diagnosed as having distal arthrogryposis type 7 (trismus-pseudocamplodactyly syndrome) based on her previously noted contractures, her recent development of trismus and a consistent family history. By age 15, the proband had experienced recurrent chest infections. She never achieved more than limited ambulation, and she normally uses a wheelchair. She has normal intellect and no evidence of cardiac or respiratory involvement.

Her family history was positive for two known additional affected individuals: the proband’s mother (age 40 years) and maternal grandmother (age 62 years) (Supplementary Fig. 1). Both had childhood onset of symptoms, including limitation of jaw opening requiring surgery in adulthood. The proband’s mother also reported ‘knock knees’, tripping while running and requiring a shoe orthotic. On examination, the proband’s mother demonstrated the praying hand sign, i.e. obligatory finger flexion on full wrist extension, most severe in the fourth digit. She and the proband’s grandmother additionally had distal contractures involving the Achilles tendons, fingers and great toe. There also was proximal joint involvement in the mother and grandmother (reportedly possibly also in a great uncle), including limited range of motion of the shoulder joint with restricted abduction (<90° in mother, and reportedly in the great uncle), extension (in the grandmother) and neck rotation (minimal in mother). The clinical characteristics are illustrated in Fig. 3 and summarized in Supplementary Table 1.

Figure 3

Clinical appearance of members of Family 2 with distal arthrogryposis type 7. Photographs from Family 2 are of the proband (Patient IV-3; A, C and E) and her mother (Patient III-3; B, D and F). Affected members of the family have limited jaw opening (trismus); A and B show their mouths fully open. Long finger flexion contractures are maximum with the wrists in full extension (pseudocamptodactyly; C and D) and less severe with the wrists in a neutral position (E and F).

Muscle biopsy is consistent with the diagnosis of nemaline myopathy

The proband and her affected mother underwent diagnostic muscle biopsies at age 5 and age 32 years, respectively. Both biopsies have scattered atrophic fibres and type I fibre predominance, but no myonecrosis or regeneration. General myopathic features including increased internal nuclei and endomysial fibrosis were present and were more severe in the mother’s biopsy (Fig. 4A and F, respectively). Numerous nemaline rods are evident on Gomori trichrome staining (Fig. 4B and G) and confirmed by electron microscopy (Fig. 4E and J). No cores are detected in the proband (Fig. 4D), whereas rare areas of pallor consistent with minicores are evident on cytochrome c oxidase staining in the mother (Fig. 4I).

Figure 4

Muscle biopsies reveal nemaline myopathy in Family 2 with distal arthrogryposis type 7. The proband’s biopsy is illustrated in A–E, whereas the proband’s mother’s biopsy is illustrated in F–J. There is increased variation in fibre size, and a small number of muscle fibres contained numerous internal nuclei (A and F, haematoxylin and eosin). The older patient also has mild endomysial fibrosis (F). Enhanced granular staining on Gomori trichrome (B and G) is shown by electron microscopy to be nemaline rods and nemaline bodies (E and J). A type I fibre predominance is present (C1 and H1, slow myosin heavy chain; C2 and H2, fast myosin heavy chain). Oxidative stains reveal no core-like structures in the proband (D), but rare minicores in the proband’s mother (I).

Targeted genetic testing identifies a novel TPM2 mutation

Based on the muscle biopsy finding, conventional Sanger sequencing was performed on DNA from the proband for the following candidate genes: MYH8 (Toydemir et al., 2006), ACTA1, CFL2, TPM3 and TPM2. A three base pair deletion (c.20_22del) in TPM2 was uncovered and subsequently confirmed to be present in the proband’s mother and grandmother. This mutation, identical to that found in Family 1, is again predicted to cause an in-frame deletion of the lysine at position 7 (p.K7del).

Family 3

Clinical and diagnostic assessment of a second family with autosomal dominant distal arthrogryposis type 7

The proband (Patient III-1) is now aged 26 years and presented at the age of 13 years with gait difficulties. She had walked late at 18–24 months, had always toe walked and had difficulties running. Examination results showed additional clinical features shared by all affected family members that included kyphosis, a short trunk and neck and limited jaw opening. She had contractures of the wrists, shoulders, hips and ankles. There was weakness of the trunk and of distal muscles, particularly in the lower limbs and particularly involving the anterior compartment of the lower limbs. Her muscle strength and function have remained stable for the subsequent 13 years. ECG is normal, and lung function shows preservation of forced vital capacity at >75% expected. Her brother (Patient III-2) had similar gait difficulties on presentation at age 9 years with similar body habitus. The father of the proband (Patient II-2) is affected and is still ambulant at age 57 years. The proband’s paternal grandmother (Patient I-2) had a stiff gait with a ‘humped back’; she died unexpectedly of respiratory failure in her fifties.

The proband has three children who are all affected: the eldest (Patient IV-1, aged 8 years) has congenital bilateral vertical talus, distal lower limb weakness and wasting, a short trunk and kyphosis. The other siblings (Patient IV-2, aged 6 years; and Patient IV-3, aged 2 years) have ankle contractures, but no talus abnormality. All three walked at ages <18 months with a stiff gait. The entire pedigree of the family is presented in Supplementary Fig. 1.

All affected individuals of this pedigree have mildly raised creatine kinase levels at two to four times the upper limit of normal. Quadriceps muscle biopsies from Patients III-1 and III-2 show type 1 fibre predominance and numerous nemaline rods, with minicore-like areas on oxidative stains (data not shown). The ultrastructure of each biopsy is illustrated in Supplementary Fig. 3. Candidate gene screening by conventional Sanger sequencing was performed on the following genes: ACTA1, RYR1, TPM3 and TPM2. As with Families 1 and 2, a 3-bp deletion (c.20_22del) in TPM2 was detected that segregates with disease in affected members of Family 3.

Family 4

Clinical and diagnostic assessment of a family with stiffness of gait and distal arthrogryposis type 7

The proband of Family 4 is a 23-year-old female subject who presented with increasing gait stiffness. Since early childhood, she was observed by her family to walk with straightened legs and to have a ‘stiff posture’. There has been gradual worsening of these symptoms over time. Currently, she reports that an increased pace of walking or stair climbing worsens the stiffness leading her legs to ‘lock-up’; this problem only resolves after a period of standing. She also reports difficulty reaching for objects with her arms, but denies any change in her fine motor skills.

The proband’s father, paternal uncle and paternal cousin all carry a diagnosis of nemaline myopathy. This diagnosis was established based on muscle biopsy of the paternal uncle (who was born with clubfoot). All affected family members remain ambulatory, and the proband’s father (aged 54 years) continues to work in construction. No affected members have a history of respiratory or cardiac problems.

Examination results of the proband (at age 23 years) and her father (at age 54 years) revealed limited opening of the jaw and pseudocamptodactyly, consistent with a diagnosis of distal arthrogryposis type 7 (Supplementary Fig. 4). High-arched palate and fifth digit clinodactyly were also documented in the proband. The proband had 4+/5 weakness in bilateral deltoids but otherwise showed normal tone and strength. The father had mild weakness in biceps, triceps and knee flexors/extensors. He had considerable wasting of the pectoralis and proximal arm muscles and has developed bilateral Achilles contractures with limited flexion of the feet. A summary of the clinical characteristics and biopsy features of Family 4 are presented in Supplementary Table 1.

Biopsy reveals nemaline myopathy and targeted gene testing identifies TPM2 mutation

The proband underwent left quadriceps muscle biopsy at the age of 23 years that revealed myofibres with wide variation in diameter including hypertrophy and atrophy (Supplementary Fig. 5). The myonuclei were mostly in a peripheral location with a moderate increase in internalized nuclei. The fibre type distribution appeared normal; both fibre types were represented among the atrophic and hypertrophic fibres in the biopsy. Gomori trichrome stain demonstrated irregular subsarcolemmal or paracentral, dark red, rod-like inclusions in numerous fibres. Electron microscopy found numerous nemaline rods and myofibrillar disarray including Z-band streaming, but no classic cores. A muscle biopsy performed in the paternal uncle (family member, Patient II-2) was not available for review. However, the original pathological diagnosis was nemaline myopathy. Based on the biopsy results, candidate gene mutation testing by Sanger sequencing was performed on the proband. No pathological sequence variants were found in RYR1 and ACTA1. TPM2 gene testing revealed the three base pair deletion (c.20_22del) identified in Families 1–3.

Western blot analysis shows no change in the distribution of tropomyosin isoforms

The tropomyosins in muscle biopsies from Family 1 and Family 4 probands were evaluated by western blot (data not shown). The amount of tropomyosin and the distribution of tropomyosin isoforms in the patient biopsies did not differ from control human skeletal muscle.

Computational modelling predicts that p.K7del disrupts TPM2 α-helical structure and adversely affects its intermolecular packing

TPM2 forms both homodimers and heterodimers and also oligomerizes into head-tail polymers (Perry, 2001; Murakami et al., 2008). We used computational modelling based on the established crystal structure for TPM2 to predict the consequences of the p.K7del mutation (Fig. 5). Based on the typical behaviour of α-helical structures, we assumed that the unchanged residues from positions 8–284 create a stable α-helix that anchors the overall TPM2 dimer. Deletion of the lysine at position 7 only displaces residues 1–5 because the lysines at positions 5 and 6 take the positions of lysines 6 and 7, respectively. Figure 5 shows the amphipathic character responsible for stabilizing the parallel α-helices of the coiled-coil motif at the N-terminus. The p.K7del mutation creates significantly reduced hydrophobic contacts on the interior of the first turn of the N-terminus. In Fig. 5, the green curves denote how large hydrophobic contacts that stabilize the wild-type become significantly smaller in p.K7del containing dimers. In the wild-type dimers, hydrophilic residues are exposed to solvent (large blue curves in Fig. 5), but the p.K7del mutation exposes hydrophobic residues that interact poorly with water and present a different surface to any protein-binding partner (dashed blue curve). Importantly, this change likely alters not only the local interaction within TPM2 but also the head-to-tail packing of the coils, as they associate on the surface of actin. Figure 5D shows how residues 1–5 of the N-termini of one dimer provide the great majority of contacts to the tail of the next dimer in the head-to-tail packed oligomer.

Figure 5

Computational modelling of p.del7K mutation in TPM2. Top: ‘Wheel diagrams’ of the N-termini of TPM2 are shown for wild-type (WT) homodimers (A), WT-p.del7K heterodimers (B), and p.del7K homodimers (C). The N-termini of the wild-type are stabilized by their amphipathic character, hydrophobic at the interface (green curves) and hydrophilic at the outer surface (blue curves). The p.del7K mutation reduces the hydrophobic contacts and exposes unfavourable residues to solvent. (D) The N-terminal region is where head-to-tail polymerization occurs, and residues 1–5 contribute much of the contacts. The helices of the tail are in red and white (protein surfaces are not shown for clarity). Residues 1–5 are highlighted on the head of the blue and black helices of the next dimer in the polymer. Altering residues 1–5 adversely affects the head-to-tail packing.

Altered p.K7del protein distribution as detected by in vivo modelling in the zebrafish

To determine the consequence(s) of the novel p.K7del mutation on TPM2 in vivo, we modelled the mutation by exogeneous expression of complementary DNA in the developing zebrafish. p.K7del-zTPM2-GFP or wild-type zTPM2-GFP were injected into 1-cell stage zebrafish embryos, and expression was examined in isolated myofibres from 4- and 7-day post-fertilization larvae. Wild-type zTPM2-GFP produced an expected pattern of GFP fluorescence that encompassed the Z-line and extended nearly to the centre of the sarcomere (Figs 6 and 7). This pattern is consistent with expression of WT-zTPM2-GFP along the extent of the thin filament, a conclusion supported by co-staining with α-actinin (to mark the Z-line; Fig. 6A) and tropomodulin (Tmod1, to mark the end of thin filament; Fig. 7A). p.K7del-zTPM2-GFP, on the other hand, was restricted almost entirely to the Z-band area, as evidenced by the overlap of p.K7del GFP fluorescence with α-actinin staining (Fig. 6B) and its clear separation from Tmod1 (Fig. 7B).

Figure 6

p.K7del-TPM2-GFP is aberrantly expressed, and it disrupts sarcomere length in the developing zebrafish. Zebrafish embryos were injected at the 1-cell stage with either wild-type zTPM2-GFP (WT TPM2) complementary DNA or zTPM2-GFP with the three base pair deletion reported in the affected patients (Δ7K TPM2). Myofibres from 7-day old larvae were isolated, immunostained for α-actinin (red), and visualized by confocal microscopy. Left: Wild-type TPM2 expressing myofibres had a thick band of GFP fluorescence (green) that spanned either side of the Z-band (labelled in red with α-actinin). This pattern is the expected pattern for TPM2. Middle: Δ7K TPM2 expression is largely confined to the region of the Z-band, displaying a much tighter band of GFP fluorescence. This observation was seen regardless of fibre size. α-actinin staining was found in aberrant accumulations throughout the myofibre. Top right: The width of GFP fluorescence was quantified. Wild-type TPM2 fluorescence is more than twice the thickness of D7 TPM2 (n = 15 fibres per condition; wild-type TPM2 = 1.56 ± 0.17 µm versus D7 TPM2 = 0.51 ± 0.09 µm; P < 0.0001). Scale bar for both columns = 2 µm. Bottom right: Sarcomere length was quantified by measuring the distance between Z-bands (as detected by α-actinin staining). Sarcomeres were significantly smaller in D7 TPM2 myofibres. (n = 10 fibres examined per condition, and 10 sarcomeres per fibre were measured. Wild-type TPM2 = 1.94 ± 0.03 µm versus D7 TPM2 = 1.41 ± 0.05 µm. P < 0.0001).

Figure 7

Δ7K-TPM2 does not span the length of the thin filament. TPM2-GFP fluorescence in zebrafish myofibres was compared with immunostaining for Tmod1. (A) Wild-type TPM-GFP (WT-TPM2-GFP) was detected in a thick band that extended to include/overlap with Tmod immunostaining (bottom left, orange staining representing co-labelling). (B) p.K7del-TPM2-GFP (Δ7K-TPM2-GFP), on the other hand, was found in a thin band that did not overlap with tropomodulin staining (bottom right). Scale bar = 2 µm.

The restricted localization of the mutant TPM2 can be represented quantitatively by measuring the width of GFP staining in myofibres expressing either wild-type or p.K7del-TPM2-GFP complementary DNA. Wild-type TPM2 has an average width of 1.56 µm, whereas p.K7del has a significantly smaller average width of 0.51 µm (Fig. 6C). Of note, overall sarcomere length, as determined by measuring the distance between α-actinin bands, is also significantly reduced in myofibres expressing p.K7del TPM2. Specific values are 1.94 µm for wild-type and 1.41 µm for p.K7del TPM2 (Fig. 5D). Overall, these findings suggest that the consequence of the p.K7del mutation is to restrict TPM dimers containing p.K7del TPM2 to the Z-band, either by inhibiting head-to-tail polymerization or by interfering with local TPM2-protein interactions. Of note, areas of peri-membranous accumulations of α-actinin (arrows) were observed in p.K7del TPM2 expressing myofibres, but not wild-type TPM2 myofibres (Fig. 6, n = 25 fibres examined per condition). Such accumulations are consistent with changes observed by α-actinin immunostaining in nemaline myopathy (Wallgren-Pettersson et al., 1995; Telfer et al., 2012), suggesting that p.K7del TPM2 expression is sufficient to cause formation of these abnormal disease-associated structures in vivo.

Discussion

TPM2 mutations have previously been reported as a rare cause of nemaline myopathy, cap myopathy, Escobar syndrome and distal arthrogryposis types 1A and 2B (Donner et al., 2002; Sung et al., 2003; Lehtokari et al., 2007; Tajsharghi et al., 2007a; Ohlsson et al., 2008; Clarke et al., 2009). Mutations have been reported throughout most of the gene, though none have yet been detected in the N-terminal region (Fig. 2A). In this study, we report a novel TPM2 mutation in association with two distinct clinical entities: core-rod myopathy and nemaline myopathy presenting as distal arthrogryposis type 7. Support for the pathogenicity of the mutation comes from several observations: (i) it causes deletion of a highly conserved amino acid; (ii) it segregates with disease in four independent pedigrees; (iii) computational modelling predicts that it will have deleterious consequences; and (iv) expression of the mutant allele in the whole organism context results in obvious alterations in TPM2 protein localization and aberrant accumulation of α-actinin. The finding of the 20_22del mutation in four separate families makes it the most common recurrent TPM2 mutation and one of the most common overall myopathy-associated mutations.

By identifying two new conditions associated with TPM2 mutations, our findings extend the clinical and histopathological spectrum of TPM2-related myopathies (recently reviewed by Tajsharghi et al., 2012). When considered in combination with other recent studies, it seems likely that TPM2 mutations are associated with a broad range of congenital myopathies and distal arthrogryposis syndromes, and that TPM2 mutations are likely to be more common than first appreciated. Our data additionally highlight the high degree of potential clinical variability associated with TPM2 mutations. Consideration of all known cases does not indicate a unifying set of shared clinical features. A commonality between the clinical features in our four families is the presence and/or early development of distal extremity contractures in combination with the development of trismus. The development of early onset distal contractures, although not described in all individuals with TPM2 mutations, is perhaps the most consistent clinical feature, and implies that TPM2, as opposed potentially to TPM1 or TPM3, plays an important role during the development of muscle and for muscle function in utero. This conclusion is in keeping with previous observations of TPM2 (but not TPM3) mutations in three syndromes characterized by congenital contractures (Tajsharghi et al., 2012): Escobar syndrome, distal arthrogryposis type 1A and distal arthrogryposis type 2B.

Examination results of the muscle biopsies from the four families revealed, among other features, the presence of nemaline rods and nemaline bodies. This is consistent with other reports of patients with TPM2-related myopathies, where either nemaline rods or caps (a morphological variant of nemaline rod/bodies) are seen in the majority of biopsies (Donner et al., 2002; Lehtokari et al., 2007; Tajsharghi et al., 2007a, b; Ohlsson et al., 2008; Clarke et al., 2009). Of note, this is the first observation of nemaline rods in a patient with a TPM2-related distal arthrogryposis syndrome. However, in previous studies, one individual with distal arthrogryposis did not have a muscle biopsy (Sung et al., 2003), and the other (who had type I fibre predominance only) was biopsied early in life (Tajsharghi et al., 2007a). It is, of course, formally possible that both individuals would develop nemaline rods over time, as our patients were biopsied at ages 5, 9 and 32 years. On the other hand, an individual with Escobar syndrome (a distinct form of arthrogryposis with multiple pterygium) and a recessive TPM2 mutation was found to have nemaline rods on a muscle biopsy obtained at the age of 2 years (Monnier et al., 2009).

The specific mutation identified in the two families reported here is novel and located in a region of the TPM2 gene where no mutation has previously been uncovered. The most N-terminal mutation previously reported was an amino acid substitution at position 41. The functions of the extreme N-terminus of TPM2 are less clearly understood than are those of the remainder of the molecule. However, several lines of evidence suggest that approximately the first dozen amino acids mediate the head-to-tail polymerization of tropomyosin polymers. These include deletion studies in vitro (Hitchcock-DeGregori and Heald, 1987) as well as crystallographic modelling (Murakami et al., 2008). In addition, studies of an N-terminal mutation of TPM3 (Met8Arg) has shown that this mutation disrupts actin-TPM3 interactions and likely impairs head-to-tail association of TPM3 along the thin filament (Moraczewska et al., 2000; Ilkovski et al., 2008). Extrapolation from studies of other tropomyosins has also implicated the extreme N-terminus in mediating interaction with tropomodulin (Kostyukova, 2008).

To explore potential pathomechanisms related to the p.7Kdel mutation, we performed both computational modelling and in vivo studies in the zebrafish. The molecular modelling suggests that p.7Kdel has the potential to disrupt both the head-to-tail dimerization and interactions with other proteins that bind to this part of TPM2. Our in vivo modelling, which revealed restriction of mutant TPM2 expression to the Z-band, suggests that the mutation interferes with TPM2 polymerization; the discontinuous distribution of TPM2 along the full length of the thin filament would be consistent with a loss of TPM2 polymers. However, at this point, we cannot exclude the possibility that the abnormal distribution of mutant TPM2 is owing to an inability to bind either actin or tropomodulin instead of an inability to polymerize. Additional studies, as well as comparisons with other previously reported TPM2 mutations predicted to alter actin binding but not polymerization, will be necessary to better distinguish between these possibilities.

Our study marks the first examination of a TPM2 mutation in a whole organism context, and suggests that the zebrafish model is well suited for the study of thin filament protein mutations, as supported by our observation of abnormal accumulations of α-actinin by immunostaining in fibres expressing p.K7del TPM2. We have previously reported similar immunostaining changes in a zebrafish model of nemaline myopathy owing to nebulin (NEB) mutation (Telfer et al., 2012), and abnormal immunostaining of α-actinin has additionally been observed in muscle from patients with nemaline myopathy (Wallgren-Pettersson et al., 1995). In the future, as we develop our chimeric p.K7del TPM2-GFP zebrafish into germline transgenics, we will be able to use the zebrafish to understand the functional impact of the mutation on contractile properties and motor behaviours.

Conclusion

In this study, we have identified a novel single amino acid deletion in the TPM2 gene responsible for two distinct clinical-pathological conditions: core-rod myopathy and nemaline myopathy presenting as distal arthrogryposis type 7 (trismus-pseudocamptolodactyly syndrome). Furthermore, we have demonstrated that this mutation impairs TPM2 function by interfering with intermolecular interactions necessary to establish TPM2 expression along the length of the thin filament.

Funding

S.A.M. is partially supported by the Iowa Wellstone Muscular Dystrophy Cooperative Research Centre (U54- NS053672). T.C. and H.J. were supported in part by the Guys and St Thomas’s Trust Charitable Foundation. J.J.D is supported by NIH 1K08AR054835 and MDA186999. This work was partially supported by a pilot grant from the Centre for Genomic Medicine and the Taubman Medical Institute at the University of Michigan.

Supplementary material

Supplementary material is available at Brain online.

Acknowledgements

The authors acknowledge the participation of the families.

Footnotes

  • *These authors contributed equally to this work.

  • Present address: Department of Neurology, University of Texas Southwestern, Dallas, TX, USA

Abbreviations
GFP
green fluorescent protein
TPM2
β-Tropomyosin

References

View Abstract