Introduction

The cardiomyopathies are heart muscle disorders associated with significant morbidity and mortality. These disorders are classified by the World Health Organization (WHO) into four forms^[1]: (1) dilated cardiomyopathy (DCM), (2) hypertrophic cardiomyopathy (HCM), (3) restrictive cardiomyopathy (RCM), and (4) arrhythmogenic right ventricular cardiomyopathy (ARVC). Recently, another cardiomyopathy, left ventricular noncompaction (LVNC), has gained attention although it has not yet attained separate classification; however, this will change in the near future.

The most common cardiomyopathy is DCM, accounting for approximately 55% of cases.^[1] The importance of these disorders lies in the fact that they are responsible for a high proportion of cases of heart failure and sudden death, as well as the need for transplantation. The mortality rate in the United States due to cardiomyopathy is greater than 10,000 deaths per annum, with DCM being the major contributor to this death rate.^[2] The total cost of health care in the United States focused on cardiomyopathies is in the billions of dollars and only limited success has been achieved.^[3,4] In order to achieve improved care and outcomes in children and adults, understanding of the causes of these disorders has been sought in earnest over the past decade.

DCM has become a popular target of research over the past 7-8 years, with multiple genes identified during that time period. These genes appear to encode two major subgroups of proteins, cytoskeletal and sarcomeric proteins.^[5,6] In this review, the underlying basis of DCM will be discussed.

In order to understand the mechanisms responsible for the development of the clinical DCM phenotype, an understanding of normal cardiac structure is necessary.

Normal Cardiac Structure

Cardiac muscle fibers are composed of separate cellular units (myocytes) connected in series,^[7] with the myocytes joined at each end to adjacent myocytes at the intercalated disc, the specialized area of interdigitating cell membrane (Fig. 1). The intercalated disc contains gap junctions (containing connexins), and mechanical junctions, composed of adherens junctions (containing N-cadherin, catenins, and vinculin) and desmosomes (containing desmin, desmoplakin, desmocollin, desmoglein). Cardiac myocytes are surrounded by a thin membrane (sarcolemma) and the interior of each myocyte contains bundles of longitudinally arranged myofibrils. The myofibrils are formed by repeating sarcomeres, the basic contractile units of cardiac muscle composed of interdigitating thin (actin) and thick (myosin) filaments (Fig. 1), which give the muscle its characteristic striated appearance.^[8] The thick filaments are composed primarily of myosin but additionally contain myosin binding proteins C, H, and X. The thin filaments are composed of cardiac actin, α-tropomyosin (α-TM), and troponins T, I, and C (cTnT, cTnI, cTnC). In addition, myofibrils contain a third filament formed by the giant filamentous protein, titin, which extends from the Z-disc to the M-line and acts as a molecular template for the layout of the sarcomere. The Z-disc at the borders of the sarcomere is formed by a lattice of interdigitating proteins that maintain myofilament organization by cross-linking antiparallel titin and thin filaments from adjacent sarcomeres (Fig. 1). Other proteins in the Z-disc include α-actinin, nebulette, telethonin/T-cap, capZ, MLP, myopalladin, myotilin, Cypher/ZASP, filamin, and FATZ.^[8,9]     

Finally, the extrasarcomeric cytoskeleton, a complex network of proteins linking the sarcomere with the sarcolemma and the extracellular matrix (ECM), provides structural support for subcellular structures and transmits mechanical and chemical signals within and between cells. The extrasarcomeric cytoskeleton has intermyofibrillar and subsarcolemmal components, with the intermyofibrillar cytoskeleton composed of intermediate filaments (IFs), microfilaments, and microtubules.^[10] Desmin IFs form a three-dimensional scaffold throughout the extrasarcomeric cytoskeleton with desmin filaments surrounding the Z-disc, allowing for longitudinal connections to adjacent Z-discs and lateral connections to subsarcolemmal costameres.^[10,11] Microfilaments composed of nonsarcomeric actin (mainly γ-actin) also form complex networks linking the sarcomere (via α-actinin) to various components of the costameres. Costameres are subsarcolemmal domains located in a periodic, grid-like pattern, flanking the Z-discs and overlying the I-bands, along the cytoplasmic side of the sarcolemma. These costameres are sites of interconnection between various cytoskeletal networks linking sarcomere and sarcolemma and are thought to function as anchor sites for stabilization of the sarcolemma and for integration of pathways involved in mechanical force transduction. Costameres contain three principal components: the focal adhesion-type complex, the spectrin-based complex, and the dystrophin/dystrophin-associated protein complex (DAPC).^[12] The focal adhesion-type complex, composed of cytoplasmic proteins (i.e., vinculin, talin, tensin, paxillin, zyxin), connect with cytoskeletal actin filaments and with the transmembrane proteins α-, β-dystroglycan, α-, β-, γ-, δ-sarcoglycans, dystrobrevin, and syntrophin.^[12-15] Several actin-associated proteins are located at sites of attachment of cytoskeletal actin filaments with costameric complexes, including α-actinin and the muscle LIM protein (MLP). The carboxy-terminus (C-terminus) of dystrophin binds β-dystroglycan (Fig. 1), which in turn interacts with α-dystroglycan to link to the ECM (via α-2-laminin). The amino-terminus (N-terminus) of dystrophin interacts with actin. Also notable, voltage-gated sodium channels and potassium channels colocalize with dystrophin, β-spectrin, ankyrin, and syntrophins while potassium channels also interact with the sarcomeric Z-disc and intercalated discs.^[16-22] Since arrhythmias and conduction system diseases (CDDC) are common in children and adults with DCM, this could play an important role. Hence, disruption of the links from the sarcolemma to ECM at the dystrophin C-terminus and those to the sarcomere and nucleus via N-terminal dystrophin interactions could lead to a "domino effect" disruption of systolic function and development of arrhythmias.

Dilated Cardiomyopathy

Clinical Aspects

Idiopathic DCM (IDCM) is characterized by increased ventricular size (i.e., left ventricular or biventricular dilation) and reduced ventricular contractility (Fig. 2) in the absence of coronary artery disease, valvular abnormalities, or pericardial disease.^[1] Mitral regurgitation is common, as are ventricular arrhythmias, particularly ventricular tachycardia (VT), torsade de pointes (TdP), and ventricular fibrillation (VF). Clinical features include the signs and symptoms of heart failure.^[23]

Although the overall incidence varies, it is believed that DCM occurs in at least 40/100,000 in the adult population^[24] but is less in children and varies by ae.^[25-27] The prevalence and incidence of DCM appear to be increasing.^[23,28] Depending on the diagnostic criteria used, the annual incidence in adults varies between 5 and 8 cases/100,000 population and is variable in chlidren^[24-27]; the true incidence is probably underestimated by these figures, since many asymptomatic cases go unrecognized. Nearly five million Americans have heart failure currently with an increasing incidence with age.^[23]

Molecular Genetics of DCM

Over the past decade, progress has been made in the understanding of the genetic etiology of familial DCM (FDCM). Initial progress was made studying families with X-linked forms of DCM, with the autosomal dominant forms of DCM beginning to unravel over the past few years.

X-Linked Cardiomyopathies

X-Linked Dilated Cardiomyopathy (XLCM)

First described in 1987 by Berko and Swift^[29] as DCM occurring in males in the teen years and early twenties with rapid progression from congestive heart failure (CHF) to death due to VT/VF or transplantation, these patients are distinguished by elevated serum creatine kinase muscle isoforms (CK-MM). Female carriers tend to develop mild to moderate DCM in the fifth decade and the disease is slowly progressive. Towbin et al. and Muntoni et al. were the first to identify the disease-causing gene as dystrophin and characterize the functional defect,^[30,31] a loss of cardiac dystrophin. Multiple mutations have been identified in dystrophin in patients with XLCM, the vast majority of mutations affecting the 5' end of the gene and affecting the muscle promoter or the N-terminus of the protein.^[32-37]

Dystrophin is a cytoskeletal protein which provides structural support to the myocyte by creating a lattice-like network to the sarcolemma.^[38] In addition, dystrophin plays a major role in linking the sarcomeric contractile apparatus to the sarcolemma and ECM.^[39-42] Furthermore, dystrophin is involved in cell signaling, particularly through its interactions with nitric oxide synthase.^[43-47] The dystrophin gene is responsible for Duchenne and Becker muscular dystrophy (DMD/BMD) when mutated as well.^[48] These skeletal myopathies present early in life (DMD is diagnosed before age 12 years while BMD is seen in teenage males older than 16 years of age) and the vast majority of patients develop DCM before the 25th birthday.^[49] In most patients, CK-MM is elevated similar to that seen in XLCM; in addition, manifesting female carriers develop disease late in life, similar to XLCM. Furthermore, immunohistochemical analysis demonstrates reduced levels (or absence) of dystrophin, similar to that seen in the hearts of patients with XLCM.

Murine models of dystrophin deficiency demonstrate abnormalities of muscle physiology based on membrane structural support abnormalities.^[50-53] In addition to the dysfunction of dystrophin, mutations in dystrophin secondarily affects proteins which interact with dystrophin. At the N-terminus, dystrophin binds to the sarcomeric protein actin, a member of the thin filament of the contractile apparatus. At the C-terminus, dystrophin interacts with α-dystroglycan, a dystrophin-associated membrane-bound protein, which is involved in the function of the DAPC, which includes β-dystroglycan, the sarcoglycan subcomplex (α, β, γ, δ, and δ sarcoglycan), syntrophins, and dystrobrevins (Fig. 1). In turn, this complex interacts with α-2-laminin and the ECM.^{[14,39,42,54]} Like dystrophin, mutations in these genes lead to muscular dystrophies with or without cardiomyopathy, supporting the contention that this group of proteins is important to the normal function of the myocytes of the heart and skeletal muscles.^{[14,39,42,55]} In both cases, mechanical stress^[56-58] appears to play a significant role in the age-onset dependent dysfunction of these muscles. The information gained from the studies on XLCM, DMD, and BMD led us to hypothesize that there is a "final common pathway" for DCM and that DCM is a disease of the cytoskeleton/sarcolemma, which affects the sarcomere.^[6,59] We have also shown that dystrophin mutations play a role in IDCM in males.^[37] Further support for this concept was provided by Badorff, Knowlton, and colleagues, who showed that LV dysfunction seen in coxsackievirus myocarditis resulted from dystrophin cleavage by the viral enzyme enteroviral protease 2A.^[60,61] Hence, acquired DCM results from disruption of dystrophin and the cardiac cytoskeleton as well.

Autosomal Dominant DCM

The most common form of inherited DCM is the autosomal dominant form of disease.^[5] These patients present as classic "pure" DCM or DCM associated with CDDC. In the latter case, patients usually present in the twenties with mild CDDC, which can progress to complete heart block over decades.^[62,63] DCM usually presents late in the course but is out-of-proportion to the degree of CDDC.^[64]

Genetic heterogeneity exists for autosomal dominant DCM with 15 loci mapped for pure DCM and five loci for CDDC.^[5,65-67] In the case of pure autosomal dominant DCM, 13 genes have been identified to date: δ-sarcoglycan, α-actinin-2, ZASP, actin, desmin, troponin T, β-myosin heavy chain, titin, metavinculin, myosin binding protein C, α-TM, MLP, and phospholamban^[68-77].

The majority of genes identified to date encode either cytoskeletal or sarcomeric proteins. In the case of cytoskeletal proteins (desmin, δ-sarcoglycan, metavinculin), defects of force transmission are considered to result in the DCM phenotype,^[6,67] while defects of force generation have been speculated to cause sarcomeric protein-induced DCM.^[72]

Desmin is a cytoskeletal protein that forms IFs specific for muscle^[10,11] and forms connections between the nuclear and plasma membranes of cardiac, skeletal, and smooth muscle. Desmin is found at the Z-lines and intercalated disk of muscle, and its role in muscle function appears to involve attachment or stabilization of the sarcomere. Mutations in this gene may also cause skeletal myopathy.

Another DCM-causing gene, δ-sarcoglycan, is a member of the sarcoglycan subcomplex of the DAPC.^[78,79] This gene encodes for a protein involved in stabilization of the myocyte sarcolemma as well as signal transduction. Mutations identified in familial and sporadic cases resulted in reduction of the protein within the myocardium.^[68] In the absence of δ-sarcoglycan, the remaining sarcoglycans (δ, β, γ, Σ) cannot assemble properly in the endoplasmic reticulum. Mouse models of δ-sarcoglycan deficiency demonstrate dilated, hypertrophic cardiomyopathy, sarcolemmal fragility, and disrupted vasculin smooth muscle that leads to vascular spasm, including coronary spasm.^[80,81] In addition, mutations in this gene lead to the phenotype of the cardiomyopathic Syrian hamster.^[82-84] Other human mutations in δ-sarcoglycan cause a form of autosomal recessive limb girdle muscular dystrophy (LGMD2F), which is rarely associated with heart disease.^[85,86]

The cytoskeletal protein-encoding gene, metavinculin, encodes vinculin and its splice variant metavinculin.^[87,88] Vinculin is ubiquitously expressed and metavinculin is coexpressed with vinculin in heart, skeletal, and smooth muscle. It is localized to subsarcolemmal costameres in the heart where they localize to subsarcolemmal costameres and interact with α-actinin, talin, and γ-actin to form a microfilamentous network linking cytoskeleton and sarcolemma. In addition, these proteins are present in adherens junctions in intercalated disks and participate in cell-cell adhesion.^[87]

Mutations in the sarcomere may produce HCM or DCM. In the latter case, abnormalities in force generation or transmission are thought to contribute to the development of this phenotype.^[72] Cardiac actin^[70] is a sarcomeric protein that is a member of the sarcomeric thin filament interacting with tropomyosin and the troponin complex. Actin links the sarcomere to the sarcolemma via its binding to the N-terminus of dystrophin.^[89] The DCM-causing mutations are believed to result in DCM by causing force transmission abnormalities.

In addition to mutations in the thin filament protein actin, mutations in the thick filament protein-encoding gene β-myosin heavy chain have been shown to cause DCM associated with sudden death in at least one infant, as well as DCM.^[72,75] Mutations in this gene are thought to perturb the actin-myosin interaction and force generation or alter cross-bridge movement during contraction. Mutations in cardiac troponin T, a thin filament protein, has been speculated to disrupt calcium-sensitive troponin C binding.^[72] Mutations in phospholamban^[77] have also been identified, which further support calcium handling as a potentially important mechanism in the development of DCM. Interestingly, Haghihi et al.^[90] identified homozygous mutations causing DCM and heart failure, while heterozygotes had cardiac hypertrophy. Recessive mutation in troponin I, which is thought to impair the interaction with troponin T, α-TM mutations have also been identified and were predicted to alter the surface charge of the protein leading to impaired interaction with actin.^[91]

A recent area of interest for evaluation at the molecular level is the Z-disc.^[92] Knoll et al.^[93] identified mutations in MLP and demonstrated that this results in defects in the interaction with telethonin.^[93] Using mouse models, they also demonstrated that MLP acts as a stretch sensor and that mutant MLP causes defects in this activity. Mohapatra et al.^[69] also demonstrated mutations in MLP in families and sporadic cases and identified abnormalities in the T-tubule system and Z-disc architecture by electron microscopy, which correlates with the histopathology seen in MLP-knockout mice.^[94] This was further supported by the finding of reduced expression of MLP in chronic human heart failure.^[95,96] In addition, mutations in α-actinin-2, which is involved in cross-linking actin filaments and shares a common actin binding domain with dystrophin, were also identified in FDCM, which disrupts its binding to MLP.^[69] Finally, Vatta et al.^[97] identified mutations in the Z-band alternatively spliced PDZ-motif protein ZASP, the human homolog of the mouse cypher gene which, when disrupted, leads to DCM.^[98] Multiple mutations in this gene were identified in families and sporadic cases of DCM and with LVNC.^[97,99] This protein, which interacts with α-actinin-2, disrupts the actin cytoskeleton when mutated. Titin, a giant sarcomeric cytoskeletal protein, contributes to the maintenance of the sarcomere organization and myofibrillar elasticity, interacts with these proteins at the Z-disc/I-band transition zone.^[100] Mutations have been identified in FDCM as well.^[73]

Genetic heterogeneity also exists for CDDC, with genes mapped to chromosomes 1p1-1q1,^[62] 2q14-21,^[101] 3p25-22,^[102] and 6q23.^[103] The only gene thus far identified is lamin A/C on chromosome 1q21, which encodes a nuclear envelope IF protein.^[63,104,105]

Lamin A/C

The lamins are located in the nuclear lamina at the nucleoplasmic side of the inner nuclear membrane, and lamin A and C are expressed in heart and skeletal muscle.^[106-108] Mutations in this gene were initially reported to cause the autosomal dominant form of Emery-Dreifuss muscular dystrophy (EDMD),^[109,110] which has skeletal myopathy associated with DCM and CDDC. It has also been found to cause a form of autosomal dominant limb girdle muscular dystrophy (LGMD1B), which is also associated with CDDC.^[111] Multiple mutations have been identified in patients with DCM and CDDC which, in some cases, had mildly elevated CK. This gene defect appears to be relatively common in patients with CDDC.^[63,105] The mechanisms responsible for the development of DCM and conduction system abnormalities are currently unknown but are becoming unraveled.^[112]

Muscle Is Muscle: Cardiomyopathy and Skeletal Myopathy Genes Overlap

Interestingly, nearly all of the genes identified for inherited DCM are also known to cause skeletal myopathy in humans and/or mouse models. Dystrophin mutations cause DMD and BMD^[55] while δ-sarcoglycan mutations cause limb girdle muscular dystrophy (LGMD2F).^[86] Lamin A/C has been shown to cause autosomal dominant EDMD^[109,110] and LGMD1B,^[111] while actin mutations are associated with nemaline myopathy.^[113] Desmin, G4.5, α-dystrobrevin, Cypher/ZASP, MLP, α-actin, titin, β-sarcoglycan mutations also have associated skeletal myopathy,^[94,113-122] suggesting that cardiac and skeletal muscle function is interrelated and that possibly the skeletal muscle fatigue seen in patients with DCM with and without CHF may be due to primary skeletal muscle disease and not only related to the cardiac dysfunction. It also suggests that the function of these muscles has a "final common pathway" and that cardiologists and neurologists should consider evaluation of both sets of muscles.^[59]

Further support for this concept comes from studies of animal models. Mutations in δ-sarcoglycan in hamsters result in cardiomyopathy^[82-84] while mutations inallsarcoglycan subcomplex genes in mice cause skeletal and cardiac muscle disease.^[78-81] Mutations in other DAPC genes as well as dystrophin in murine models also consistently demonstrate abnormalities of skeletal and cardiac muscle function.^[50,52] Arber et al.^[94] also produced a mouse deficient in MLP, and the resultant mice develop severe DCM, CHF, and disruption of cardiac myocyte cytoskeletal architecture. Murine mutations in titin,^[123] cypher,^[98]α-dystrobrevin,^[118] desmin,^[124] and other all demonstrate cardiac and skeletal muscle disease. Finally, Badorff et al.^[60,125] has shown that the DCM that develops after viral myocarditis has a mechanism similar to the inherited forms. Using coxsackievirus B3 (CVB3) infection of mice, the authors showed that the CVB3 genome encodes for a protease (enteroviral protease 2A), which cleaves dystrophin at the third hinge region of dystrophin, resulting in force transmission abnormalities and DCM. In addition, Xiong et al.^[61] showed that abnormal dystrophin increases susceptibility to viral infection and resultant myocarditis. N-terminal dystrophin is reduced or absent in hearts of patients with all forms of DCM (ischemic, acquired, genetic, idiopathic) and reduction of mechanical stress by use of left ventricular assist devices (LVADs) results in reverse remodeling of dystrophin and of the heart itself.^[126,127]

Role of Genetic Analysis in DCM: When to Perform It and How to Use the Results for Patient Management

Now that the genes for DCM are rapidly being elucidated, an obvious next step would be to make these genetic analyses available for clinical use. When new genes for any disease are first discovered, the availability for patient and family screening is limited to research analysis by research laboratories. In general, research laboratories are typically slow to complete individual samples and not necessarily focused on providing the completed data back to the referring clinician or patient in a timely fashion and, in many cases, are forbidden to provide this information. Only when these tests move to a fee-for-service laboratory does the process change to a "patient-friendly," rapid turn-around scenario. Our diagnostic laboratory (http://www.bcm.edu/pediatrics/welsh), for instance, provides a report back to the referring physician within 3 weeks of sample receipt. This approach is revolutionizing the use of this information. Another interim approach that has been used by our research laboratory and others is the collaboration with a CLIA-approved fee-for-service laboratory in which the result from the research laboratory is re-tested by the fee-for-service laboratory and then reported to patient or physician. This must, however, be agreed to during the initial consenting process (and approved by the local Institutional Review Board, IRB).

Once the mutation analysis information is provided, how should it be used? We support the use of genetic counselors in discussing the results, defining the genetic abnormality, outlining the genetic risk to the affected subject and gene carriers, the potential theoretic genetic risk throughout carriers, and the potential theoretic genetic risk throughout the family. Currently, no genotype-phenotype data exist that will allow clinicians to risk-stratify patients based on genotype. For this reason, these tests should not be used to define management at present in affected patients but should alert clinicians to define early follow-up strategies in carriers.

Therefore, the role of genetic analysis at current is in a state of flux. Presently, it should be used to evaluate whole families for risk and should be considered at the time of diagnosis of the proband. Family members should undergo clinical evaluation initially (nuclear family) and once the gene defect is defined in the proband, family screening would be appropriate. In the future, when all genes responsible for DCM are known, a more robust approach will become the standard, which will include genetic screening at clinical diagnosis.

Summary

DCM occurs due to disruption of the cytoskeleton and its linkage to the sarcomere and nucleus. Treatment approaches in the future should consider targeting these proteins and the secondary abnormalities caused by mechanical stress. Once these approaches are adopted, new studies in the treatment of these disorders will be possible.

CHR Locus	Gene	Protein
Xp21.2	DYS	Dystrophin
Xq28	G4.5	Tafazzin
1q21	LMNA	Lamin A/C
1q32	TNNT2	Cardiac Troponin T
1q42–43	ACTN	a-Actinin
2q31	TTN	Titin
2q35	DES	Desmin
5q33	SGCD	δ-Sarcoglycan
6q22.1	PLN	Phospholamban
10q22.3–23.2	ZASP/Cypher	ZASP
10q22–q23	VCL	Metavinculin
11p11	MYBPC3	Myosin Binding Protein C
11p15.1	MLP	Muscle LIM Protein
14q12	MYH7	β-Myosin Heavy Chain
15q14	ACTC	Cardiac Actin
15q22	TPM1	α-Tropomyosin

 

Jeffrey A. Towbin, M.D., Professor & Chief, Pediatric Cardiology, Baylor College of Medicine, Texas Children's Hospital, 6621 Fannin Street, MC 19345-C, Houston, TX 77030. Fax: (832) 825-5921; E-mail: jtowbin@bcm.tmc.edu