Article Text


Molecular biology and genetics
Idiopathic restrictive cardiomyopathy in children is caused by mutations in cardiac sarcomere protein genes
  1. J P Kaski1,
  2. P Syrris2,
  3. M Burch1,
  4. M-T Tomé-Esteban1,
  5. M Fenton1,
  6. M Christiansen3,
  7. P S Andersen3,
  8. N Sebire4,
  9. M Ashworth4,
  10. J E Deanfield1,
  11. W J McKenna1,
  12. P M Elliott1
  1. 1
    Inherited Cardiovascular Diseases Unit, Cardiac Unit, Institute of Child Health, University College London and Great Ormond Street Hospital for Children, London, UK
  2. 2
    Department of Medicine, University College London Hospital, London, UK
  3. 3
    Department of Biochemistry, Statens Serum Institute, Copenhagen, Denmark
  4. 4
    Department of Histopathology, Great Ormond Street Hospital for Children, London, UK
  1. Dr P M Elliott, The Heart Hospital, 16–18 Westmoreland Street, London W1G 8PH, UK; pelliott{at}


Background: Restrictive cardiomyopathy (RCM) is rare in childhood, but has a grave prognosis. The cause of disease in most cases is unknown.

Objective: To determine the prevalence of sarcomere protein gene mutations in children with idiopathic RCM.

Methods: Twelve patients (9 female, mean age 5.1 years) with idiopathic RCM referred between 1991 and August 2006 underwent detailed clinical and genetic evaluation. Nine had received cardiac transplants at the time of the study. The entire coding sequences of the genes encoding eight cardiac sarcomere proteins and desmin were screened for mutations. Familial evaluation was performed on first-degree relatives.

Results: Four patients (33%) had a family history of cardiomyopathy: RCM (n = 2); dilated cardiomyopathy (n = 1) and left ventricular non-compaction (n = 1). Sarcomere protein gene mutations were identified in four patients (33%): 2 in the cardiac troponin I gene (TNNI3) and 1 each in the troponin T (TNNT2) and α-cardiac actin (ACTC) genes. Two were de novo mutations and 3 were new mutations. All mutations occurred in functionally important and conserved regions of the genes.

Conclusions: Sarcomere protein gene mutations are an important cause of idiopathic RCM in childhood. We describe the first mutation in ACTC in familial RCM. The identification of RCM in a child should prompt consideration of sarcomere protein disease as a possible cause and warrants clinical evaluation of the family.

Statistics from

Restrictive cardiomyopathy (RCM) is characterised by increased stiffness of the myocardium that causes pressure within the ventricle to rise precipitously with only small increases in volume. By definition, ventricular diastolic volumes are usually normal or reduced, wall thickness is normal or mildly increased and systolic function is preserved.1 2 In childhood, RCM is rare, accounting for only 2–5% of all paediatric cardiomyopathies,3 4 but its prognosis is grave with a 2-year survival of <50%, largely owing to rapid progression of pulmonary vascular disease.515

About 30% of children with RCM have a family history of cardiomyopathy.6 8 10 12 Rare examples of familial RCM in children associated with desmin gene mutations16 or musculoskeletal abnormalities17 18 have been described, but the cause of disease in most patients is unknown. We hypothesised that RCM in children may be a disease of myofibrillar proteins. The aim of this study was therefore to determine the prevalence of sarcomere protein gene mutations in a cohort of children presenting with idiopathic RCM.



Between 1991 and August 2006, 33 children (age ⩽16 years) with RCM were assessed at Great Ormond Street (1991–2006) and Harefield Hospital (1991–2001) transplant units, London, UK. The diagnosis of RCM was based on (a) echocardiographic features of RCM (fig 1A), including atrial dilatation in the presence of normal or near-normal left ventricular (LV) cavity size (LV end-diastolic dimension z-score ⩽3) and systolic function (fractional shortening ⩾25%)1 2 9 19; (b) a maximal LV wall thickness z-score ⩽3; (c) no evidence of constrictive pericarditis and (d) raised ventricular end-diastolic pressures at cardiac catheterisation.13 Restrictive mitral inflow pulsed-wave Doppler velocities (fig 1B) and ECG features of RCM (fig 1C) provided supportive evidence.

Figure 1 Echocardiographic and ECG features of restrictive cardiomyopathy. (A) Apical four-chamber view from individual H1685 showing severe biatrial dilatation in the presence of normal left ventricular diastolic dimensions and wall thickness. (B) Pulsed-wave Doppler mitral inflow filling pattern from individual H1647 demonstrating restrictive left ventricular physiology (high E/A ratio, rapid E deceleration time and reduced isovolumic relaxation time). (C) ECG from individual H906 showing biatrial enlargement, ST-segment depression in the inferior and lateral precordial leads and Q waves in the inferior leads.

Seven patients who had died before the study were excluded from genetic evaluation. All remaining patients who were alive at the beginning of the study (March 2005) were screened for inclusion. One patient declined to participate. After review of clinical data, a further 13 patients were excluded: one patient with Noonan syndrome; four with peripheral neuropathy; three with congenital heart disease; one with LV hypertrabeculation and reduced fractional shortening on echocardiography; one with dilated cardiomyopathy (DCM) and restrictive physiology and three with LV hypertrophy (body surface area-corrected maximal LV wall thickness z-scores >3) and restrictive physiology. The final study cohort comprised 12 patients with idiopathic non-syndromic RCM, including individual H906.1 reported on previously20 (fig 2). The clinical outcome after transplantation of five of these patients has been previously reported.9

Figure 2 Breakdown of the group of patients with restrictive cardiomyopathy (RCM) at Great Ormond Street Hospital and Harefield Hospital 1991–August 2006. ASD, atrial septal defect; DCM, dilated cardiomyopathy; LV EFE, left ventricular endocardial fibroelastosis; LVH, left ventricular hypertrophy.

Clinical evaluation

For patients diagnosed before March 2005, a retrospective case note review was performed. Initial 12-lead ECG, echocardiograms and cardiac catheterisation data were reviewed when available. Patients referred after March 2005 underwent systematic evaluation, including ECG, two-dimensional, M-mode and Doppler echocardiography and cardiac catheterisation. LV end-diastolic dimensions, wall thicknesses and left atrial dimensions are expressed as a standard deviation from the body surface area-corrected mean (z-score) based on previously published normal values.21

Information on family history was obtained from medical notes, through interviews with family members and by contacting general practitioners and hospital doctors. All first-degree relatives of probands in the study were invited to undergo clinical evaluation, including ECG and echocardiography.

DNA extraction and mutation analysis

Genomic DNA was extracted using a QIAGEN kit (Qiagen, Hilden, Germany) from peripheral blood samples. Using a combination of direct sequencing and single-strand conformational polymorphism (SSCP), the protein-coding sequences of the following sarcomeric genes were screened in a collaborative effort between University College London (UCL) and Statens Serum Institut (SSI, Denmark): Troponin I (TNNI3, 8 exons), β-myosin heavy chain (MYH7, 3–40 exons), myosin-binding protein C (MYBPC3, 2–35 exons), troponin T (TNNT2, 16 exons), α-tropomyosin (TPM1, 9 exons), regulatory myosin light chain (MYL2, 7 exons), essential myosin light chain (MYL3, 6 exons) and α-cardiac actin (ACTC, 6 exons). In addition, the entire coding sequence of the desmin gene (DES, 9 exons) was screened by direct sequencing, as mutations in this gene are associated with RCM in children.16 22 All exons were amplified by PCR using primers and optimised protocols determined at UCL or SSI and scanned by fluorescent capillary array electrophoresis SSCP with the use of an ABI 3100 genetic analyser and Genotyper 3.7 software (Applied Biosystems, Foster City, California, USA).23 24 This technique has a sensitivity of 95% and a specificity of 97% for detecting mutations in hypertrophic cardiomyopathy (HCM).24 The amplified products were purified with the GFX-96 PCR purification kit (Amersham Biosciences, Piscataway, New Jersey, USA). Direct sequencing of purified PCR products was performed using the ABI BigDye Terminator sequencing kit and an ABI 3130 genetic analyser (Applied Biosystems). Sequences were analysed with Seqscape 2.5 software. Following the same methodology, a cohort of 100 unrelated healthy volunteers served as controls for every change in DNA sequence found in patients with RCM.


The study was approved by the local research ethics committee and informed written consent was obtained from all participants aged ⩾16 years or from the parents of those aged <16 years.

Statistical analysis

SPSS (version 11.0) was used for all statistical analyses. Normally distributed data are expressed as mean (95% CI); non-parametric data are expressed as median (interquartile range (IQR)). Differences between means were compared using the Student t test or one-way analysis of variance. The Fisher exact test was used for comparison of categorical data. The Kruskal–Wallis test was used to analyse non-parametric continuous data. A p value <0.05 was considered significant.


Clinical characteristics of probands

Table 1 shows the clinical characteristics of probands. Median age at diagnosis was 5.1 years (range 11 months to 14.7 years; IQR 3.4–7.9 years); nine patients (75%) were female. Nine patients (75%) were diagnosed during investigation for cardiovascular symptoms and three (25%) after screening for familial cardiomyopathy. Ten patients (83.3%) had cardiac symptoms at diagnosis. Eleven patients (91.7%) were in sinus rhythm; one patient was in atrial flutter when diagnosed with RCM at the age of 14.

Table 1 Clinical data of probands with idiopathic restrictive cardiomyopathy

Complete transthoracic echocardiography data were available in 10 patients (83.3%). The original echocardiograms for the remaining two patients were unavailable for review, but were reported as showing atrial dilatation with normal ventricular dimensions and wall thickness. Where available, mitral inflow Doppler velocities showed evidence of restrictive filling or impaired relaxation. All patients had raised LV end-diastolic pressure or pulmonary capillary wedge pressure at cardiac catheterisation (>15 mm Hg) and eight (66.7%) had raised pulmonary vascular pressures.

Histology of explanted hearts

Histopathology from explanted hearts was available in seven of nine patients (77.8%) who underwent cardiac transplantation (fig 3). Interstitial fibrosis and myocyte hypertrophy were present in all specimens. Myocyte disarray was present in three cases (42.9%). This was associated with dysplastic changes in the intramural coronary arteries in one case. Two cases showed myocyte vacuolation. One case showed hypereosinophilic inclusion bodies within the myofibres, which did not stain with periodic acid–Schiff reagent, Luxol fast blue, Congo red, or for desmin on immunohistochemistry (fig 3C). Identical findings were found in the explanted heart of this person’s sister, who died after cardiac transplantation for RCM (data not included in this report). There was no clinical evidence of skeletal muscle weakness in any family members.

Figure 3 Histopathological features of idiopathic childhood restrictive cardiomyopathy. (A) Macroscopic specimen from individual H1684 showing a non-hypertrophied, non-dilated left ventricle with a severely dilated right atrium (the left atrium, mitral valve, interatrial septum and crest of the ventricular septum have been removed for valve harvesting); (B) a section of left ventricular myocardium from individual H1685 showing myocyte disarray and interstitial fibrosis; (C) a section of left ventricular myocardium from individual H1647 showing hypereosinophilic inclusion bodies within the myofibres. These inclusions did not stain with periodic acid–Schiff reagent, Luxol fast blue or Congo red and also did not stain for desmin on immunohistochemistry.

Family screening

Pedigree analysis

Four patients (33.3%) had a family history of cardiomyopathy (RCM, n = 2; DCM, n = 1; LV non-compaction (LVNC), n = 1), with first-degree relatives affected in three cases and second-degree relatives affected in one. Of the 12 patients, two (16.7%) also had a family history of premature sudden cardiac death: in one case this was an adult relative with RCM; in the other, the proband’s brother died suddenly aged 16 years and postmortem examination showed LVNC; their father had a sudden cardiac death aged 36, but no autopsy was performed.

Clinical evaluation

Clinical screening was carried out in first-degree relatives of all probands. Thirty-three of 37 live first-degree relatives (89%) underwent clinical evaluation. In one family with a known history of RCM, two previously undiagnosed, asymptomatic subjects fulfilled diagnostic criteria for RCM. In a second family with a history of DCM, a further relative had features of DCM with restrictive LV physiology. No new cases of RCM were identified in families without a previously known history of cardiomyopathy.

Genetic investigation

Sarcomere protein gene mutations were identified in four patients (33.3%): two patients carried mutations in TNNI3 (including the previously reported patient20); one patient carried a mutation in TNNT2 and one patient carried a mutation in ACTC. An intronic sequence change in MYBPC3 was identified in a fifth patient, but its significance is uncertain. No mutations were identified in DES. Nucleotide numbering for mutations is according to genomic sequences at (accessed 24 June 2008).

TNNI3 mutations

The TNNI3 sequence change in subject H906.1 has been reported previously.20 It is a g.4792A→G transition in exon 7, which results in a Lys178Glu amino acid substitution. This was a de novo mutation, absent from both parents, and paternity was confirmed after haplotype analysis.20

Individual H1685.1 had a novel deletion of two nucleotides (g.4789_4790delAA) in exon 7 of TNNI3. This resulted in a frame shift and the introduction of a premature termination codon at amino acid position 209 (Glu177fsX209). A deletion in TNNI3 has never been described in association with RCM. The mutation was not present in her parents (paternity was confirmed).

TNNT2 mutation

Mutation analysis of TNNT2 in individual H1703.1 revealed a nucleotide substitution in exon 10 (g.9718G→A), which led to a Glu136Lys amino acid substitution. The mutation was identified in his father and brother (and not in the mother), but clinical investigations in these relatives were normal. This is a new mutation, but its true pathogenicity remains to be elucidated, given that it is present in two relatives without clinical disease and histology from the proband’s explanted heart showed myocyte vacuolation rather than the disarray commonly associated with troponin T cardiomyopathy.25

ACTC mutation

Mutation analysis in individual H1710.1 identified a nucleotide substitution in exon 5 of ACTC (g.4642G→C) that led to an Asp313His amino acid substitution. This is a new mutation. The proband’s father died after transplantation for DCM and her sister was diagnosed with a mixed RCM/DCM phenotype during family screening (LV end-diastolic dimension z-score +4, restrictive LV physiology, fractional shortening 40%, interventricular septum z-score +0.8, LV posterior wall z-score +2.3). Tissue from her father’s explanted heart was unavailable for genetic analysis and her sister refused genetic screening. However, pedigree analysis suggests that this mutation was inherited and is very likely to be disease causing (see below).

MYBPC3 sequence variation

Individual H1405.1 was found to harbour a C→T nucleotide transition in intron 3 of MYBPC3 (IVS3-82C>T). This sequence change has not been reported previously, but falls well inside the intron and is likely to be a polymorphism. Functional RNA studies are needed to confirm this.

Pathogenicity of mutations

All sequence changes identified in this study were found in patients of Caucasian descent. We consider them to be disease causing for the following reasons: (a) the appearance of de novo mutations in individuals H906.1 and H1685.1; (b) the absence of these sequence variations in 200 chromosomes from Caucasian control subjects and (c) the location of all mutations in functionally important and conserved regions of the genes. The mutations identified in TNNI3 affect highly conserved amino acids among different species and are predicted to disrupt the actin-binding domain of the troponin I protein. Similarly, the glutamic acid at position 136 in TNNT2 is highly conserved and mutation Glu136Lys would affect the troponin T–tropomyosin-binding domain. Mutation Asp313His in ACTC alters a highly conserved aspartate at position 313 and is located at an important tropomyosin-binding site.26

Comparison of sarcomeric and non-sarcomeric patients

Table 1 shows the clinical features of the patients with sarcomeric and non-sarcomeric RCM. There were no significant differences in the clinical, echocardiographic, histopathological and haemodynamic features between the two groups (data not shown).

Clinical outcome

Ten patients (83.3%) were alive at the end of the study. Nine patients (75%) had undergone cardiac transplantation. Two patients (16.7%) were alive without transplantation. One patient (8.3%) died while on the active transplant list and one died after cardiac transplantation.


This study demonstrates that very young children with RCM can have sarcomeric protein gene mutations. The findings extend the current knowledge of the distribution of sarcomere protein gene mutations and provide new insights into the complex pathophysiology of RCM, with important implications for the management of families.

Definition of RCM

There have been many attempts to define RCM,1 2 19 27 28 but consistent application of most published definitions is often difficult in everyday clinical practice. In general, all patients with RCM should have restrictive LV physiology, in which increased myocardial stiffness causes ventricular pressure to rise precipitously with only small increases in volume. This phenomenon can, however, occur in association with different ventricular morphologies including HCM, DCM and LVNC. In the WHO2 (and more recent AHA19) classification of cardiomyopathies, RCM is defined by restrictive ventricular physiology associated with normal or reduced diastolic volumes (of one or both ventricles), normal or near-normal systolic function and normal or only mildly increased ventricular wall thickness. However, even this description is problematic because of the use of subjective qualifiers such as “mild” and “near normal”. The findings in this and other studies demonstrate that, in reality, RCM is not a single entity, but is instead a heterogeneous group of disorders that can present with a spectrum of cardiac phenotypes (including HCM in family members). Furthermore, although sarcomeric protein gene mutations are a cause of HCM, RCM and DCM, they are not their defining feature. Therefore, the terms hypertrophic and restrictive cardiomyopathy refer not to specific diseases, but are instead purely descriptive terms used to characterise myocardial disease associated with a broad spectrum of genetic syndromes or systemic diseases.

Relationship between sarcomeric protein disease and restrictive LV physiology

Cardiomyocyte contraction is dependent on intracellular calcium concentration and is regulated by the troponin complex, comprising troponin I, troponin T and troponin C. Troponin I binds to actin-tropomyosin and prevents muscle contraction by inhibiting actomyosin activity. This effect is reversed by calcium binding to troponin C. This triggers a series of events including binding of the myosin head to actin and of ATP to myosin, displacement of the myosin head along the thin filament and ATP hydrolysis, resulting in force generation. Recent in vitro studies have shown that RCM-causing mutations in TNNI3 show a greater increase in Ca2+ sensitivity than HCM-causing mutations, resulting in more severe diastolic impairment and potentially accounting for the RCM phenotype in humans.29 30 The two TNNI3 mutations identified in this study fall within a functionally important region that binds to actin and increases the inhibitory effect of troponin I. The TNNT2 mutation identified in this study falls within the amino terminal of the gene, a region that is essential for filament and sarcomere assembly. In vitro studies have shown that mutations in this region result in reduced tropomyosin binding to actin.31 Recently, the first report of infantile RCM caused by a TNNT2 mutation in this region was published.32 A mutation at position 141 of TNNT2 has also been reported in association with familial DCM.33 Altered calcium sensitivity or other environmental factors may explain the restrictive phenotype in our patient, rather than the expected DCM phenotype. The histology of the explanted heart showed myocyte vacuolation rather than the typical disarray seen with troponin T disease,25 and two genetically affected relatives did not have clinical disease. Therefore, although this TNNT2 mutation is likely to be disease causing, as it occurs in a highly conserved and functionally important region of the gene and was absent from control DNA, it should be considered of uncertain significance, pending further mutation screening and functional analyses.

The ACTC mutation identified in this study falls within an important tropomyosin-binding site, in the immobilised end of the actin filament. This region is involved in force propagation, and the mutation identified may share a pathogenic mechanism with cytoskeletal protein gene mutations that cause DCM. Although we were unable to provide evidence of cosegregation in this study, a mutation a single amino acid away from this mutation has been reported in association with DCM,34 and in our study, the proband’s father underwent transplantation for DCM and her sister was found to have DCM with restrictive LV physiology on clinical screening. The reasons for the RCM phenotype in the proband are unclear.

De novo mutations

Previous reports have suggested that de novo mutations in the sarcomere protein genes are associated with a younger onset and more severe disease expression.20 35 36 Our data support this observation, in that 50% of the sequence changes identified were de novo mutations. De novo mutations may occur as a result of germline mosaicism,36 although other mechanisms have been postulated, including somatic mutations during embryogenesis affecting solely the heart progenitor cells.37

Histological findings

Where cardiac histology was available, the microscopic appearances were identical to familial HCM in over 40% of cases, including the two patients with TNNI3 mutations. One exception was a family with inclusion bodies within the myofibres (fig 1C); similar findings have been reported in 63 patients with myofibrillar myopathy.38 Although all patients in that report had muscle weakness on clinical examination, the mean age at presentation was 54 years and only one presented before the age of 10. The family in our study may represent an unusual form of this condition, with predominantly cardiac manifestations and subclinical or absent muscle pathology.

Non-sarcomeric RCM

The cause of disease in the remaining two-thirds of the patients in this study remains unknown. Given the phenotypic overlap between RCM and other cardiomyopathies, it is possible that some cases of RCM are caused by mutations in genes encoding cytoskeletal or nuclear envelope proteins, more commonly associated with DCM. Also, as with HCM, some cases of RCM may be associated with inborn errors of metabolism or storage disorders with predominantly cardiac involvement. Further studies to determine this are required.


As in previous studies on childhood RCM, this study is limited by its retrospective design. The difficulties in applying published diagnostic criteria for RCM in everyday clinical practice means that a number of patients with mild left ventricular hypertrophy who were diagnosed and treated clinically as having RCM have been excluded, resulting in an underestimation of the prevalence of sarcomeric protein disease in this population. Review of the published literature suggests that many similar patients are included in studies of childhood RCM, again illustrating the nosological confusion associated with this clinical entity.


This study shows that mutations in sarcomere protein genes are an important cause of apparently idiopathic RCM in childhood and reports the first case of familial RCM caused by a mutation in ACTC. These findings provide important clinical lessons and insights into the causes of RCM in children. Further studies to evaluate the functional and mechanistic significance of sarcomere protein gene mutations in RCM are required. The identification of RCM in a child should prompt consideration of sarcomere protein disease as a possible cause and clinical evaluation of the family.


We are grateful to the families who took part in the study. We are also grateful to Sharon Jenkins (genetic counsellor), the specialist nurses in the transplant unit at Great Ormond Street Hospital and the laboratory staff at the UCL and SSI.


View Abstract


  • Competing interests: None.

  • Funding: The study was funded by The British Heart Foundation (project grant PG/05/021). JED, PS and JPK were supported by grants from the British Heart Foundation.

  • Ethics approval: The study was approved by the local research ethics committee.

Request Permissions

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.