Introduction

Cardiomyopathy is a heterogeneous disease caused by functional abnormality of cardiac muscle and classified into primary cardiomyopathy and secondary cardiomyopathy.1 Secondary cardiomyopathy is defined as cardiomyopathy caused by extrinsic factors including ischemia, hypertension and metabolic disorders. On the other hand, diagnosis of primary cardiomyopathy is based on the exclusion of secondary cardiomyopathy and there are several different clinical types.2, 3 Hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM) are two major clinical types of primary cardiomyopathy. HCM, a major cause of sudden death in young and heart failure, is characterized by left ventricular hypertrophy, often asymmetric, accompanied by myofibrillar disarrays and diastolic dysfunction of cardiac ventricles. In contrast, DCM is characterized by dilated ventricular cavity with systolic dysfunction. Clinical symptom of DCM is heart failure and often associated with sudden death. In addition, there are other clinical types of cardiomyopathy. Restrictive cardiomyopathy (RCM) is accompanied by increased stiffness of the myocardium with diastolic dysfunction without significant hypertrophy,4 and arrhythmogenic right ventricular cardiomyopathy (ARVC) is characterized by a dilated dysfunctional right ventricle (RV), ventricular arrhythmias and fibrofatty replacement of the RV.1 Another cardiomyopathy is left ventricular noncompaction (LVNC) characterized by trabeculations in the left ventricle (LV), as well as LV hypertrophy and/or dilation.1

The etiology of primary cardiomyopathy had been unknown, but various genetic abnormalities associated with the cardiomyopathy have recently been unraveled. More than half of HCM patients have family history of the disease consistent with autosomal dominant genetic trait.5 In the case of DCM, about 20–35% patients had family history of the disease, mainly consistent with the autosomal dominant inheritance, although some familial cases can be explained by autosomal recessive or X-linked recessive trait.6, 7 Because the presence of family history suggests the genetic etiology, linkage studies have been performed to identify the disease loci in each multiplex family with cardiomyopathy. Identification of the disease loci has enabled to decipher the disease-linked mutations in the genes located within the loci. Subsequently, candidate gene approaches, focused on the genes encoding for proteins related or interacting with products of the previously identified disease genes, have been successful in unraveling novel disease genes. As shown in Table 1, many different disease genes have so far been identified. The most important issue is the overlapping of disease genes for different clinical types.

Table 1 Disease genes for hereditary cardiomyopathy
Full size table

Mutations in genes for contractile elements in cardiomyopathy

Identification of a missense mutation in cardiac β-myosin heavy chain gene (MYH7) linked to HCM in a large multiplex family was the first demonstration of the disease gene for HCM.8 At the same time an accompanying paper reported a fusion gene of MYH7 and cardiac α-myosin heavy chain gene (MYH6) in another multiplex family.9 It was later revealed that there was a linked MYH7 missense mutation10 in the family with the MYH6-MYH7 fusion gene, suggesting that MYH7 mutations are the major cause of HCM. Many investigators have analyzed HCM patients for mutations in MYH7 and many different missense mutations were identified,10 but the frequency of MYH7 mutations in the HCM patients was less than half and there were many families not linked to the MYH7 locus.10, 11, 12, 13 Linkage studies in such non-MYH7-linked HCM families have revealed mutations in α-tropomyosin gene (TPM1), cardiac troponin T gene (TNNT2) and cardiac myosin binding protein-C gene (MYBP3) as the causes of HCM. Because these disease genes encode contractile elements of sarcomere, genes for other contractile components were analyzed and lead to the identification of HCM-associated mutations in ventricular myosin essential light chain gene (MYL3), ventricular myosin regulatory light chain gene (MYL2), cardiac troponin I gene (TNNI3), cardiac α-actin gene (ACTC1) and cardiac troponin C gene (TNNC1).3, 11, 12, 13

We have reported that contractile element mutations are found in about 40% of Japanese patients with familial HCM in the heterozygous state, consistent with the autosomal dominant inheritance, and about 17%, 14% and 11% of HCM patients carried mutations in MYH7, MYBPC3 and TNNT2, respectively, while a few cases had mutations in other component genes such as MYL2, MYL3 and TNNI3 (Table 2). A systematic survey of mutations in consecutive 112 Japanese familial HCM patients showed that a few patients (two cases, 1.8%) had mutations in two different disease genes (double mutation case).14 The double mutation cases developed cardiac hypertrophy earlier than the single mutation case in the same family.14 In addition, we also reported that homozygous patients showed severer clinical manifestations than heterozygous patients in the same family, demonstrating the gene dose effect of mutation.15 Disease-related mutations can also be found in sporadic cases who had no family history (Table 2). Mutations in sporadic cases included de novo mutation cases,16, 17 but the other sporadic cases were probably due to the low penetrance of mutation, because most of the mutations found in the sporadic HCM patients were found in other familial HCM patients.

Table 2 Frequencies of disease-associated mutations in Japanese HCM
Full size table

Mutations in genes for contractile elements have also been found in DCM patients. Identification of cardiac α-actin gene (ACTC1) mutations was the first discovery of genetic cause of autosomal dominant DCM.18 Subsequently, ACTC1 mutation was found in HCM,19 first demonstration of an overlapping of disease genes for different primary cardiomyopathies. Molecular basis of different phenotypes caused by ACTC1 mutations was explained by the difference in which functional domains were affected; DCM-associated mutations were found at the α-actinin interacting domain,18 whereas HCM-associated mutations were at the interacting domain to myosin heavy chain.19 On the other hand, recent data suggest that there is a difference in folding property of actin between the DCM-associated mutation and HCM-associated mutation.20 Another example of overlapping disease gene was the identification of TNNT2 mutation in DCM.21 As discussed later, functional studies of TNNT2 mutations clearly demonstrated the difference between the DCM-associated mutation and HCM-associated mutation.22, 23 Therefore, contractile element mutations can be found in both HCM and DCM, but difference in the functional alterations may determine the different phenotypes.24 However, as shown in Table 3, frequencies of contractile element abnormalities in DCM patients were considerably small as compared with those in HCM patients.

Table 3 Frequencies of disease-associated mutations in Japanese DCM
Full size table

Initial analysis of functional changes caused by the MYH7 mutations demonstrated that contractile power generation was decreased in the presence of mutant myosin heavy chains25 and HCM-related mutations were found in other contractile elements, troponin T and α-tropomyosin. From these observations, it was hypothesized that HCM was the disease of sarcomere and the cardiac hypertrophy was a compensation of decreased contractility of cardiac muscle.26 However, the discovery of HCM-associated TNNI3 mutations at the contraction inhibitory domain16 implied that the decreased power might not be a common functional change caused by the contractile element mutations. Indeed, subsequent functional analyses of gene mutations in other contractile elements than MYH7 have revealed that most of the HCM-associated mutations resulted in an increased Ca2+ sensitivity of muscle contraction.23, 27, 28, 29, 30, 31 Because a MYH7 mutation that caused HCM in transgenic mice also increased Ca2+ sensitivity at the muscle fiber level,32 a common functional alteration due to the HCM-related sarcomere mutations may be the increased Ca2+ sensitivity. Muscle contraction is regulated by the concentration of intracellular Ca2+ that is released from sarcoplasmic reticulum via ryanodine receptor (RyR2) and restored to sarcoplasmic reticulum via sarcoplasmic reticulum Ca2+-ATPase (SERCA). When the concentration of Ca2+ is increased or decreased, muscle is contracted or relaxed, respectively. In other words, the increased Ca2+ sensitivity means a leftward shift of Ca2+-tension curve; more tension is generated by mutant contractile elements than normal elements at the same Ca2+ concentration (hyper-contraction) or muscle with mutant elements is under less relax states (diastolic dysfunction) than the normal muscles. This is consistent with the finding that characteristic features of HCM are hyper-contraction and diastolic dysfunction.

On the other hand, it was demonstrated in model mice carrying cardiomyopathy-associated TNNT2 mutations that Ca2+ sensitivity was increased by a HCM-associated delGlu160 mutation, whereas it was decreased by a DCM-associated delLys210 mutation.22 These observations indicated that DCM-associated contractile element mutations lead to systolic dysfunction of the heart. It should be noted here that drugs or chemicals that increased Ca2+ sensitivity could be effective in preventing heart failure in the mouse models carrying the DCM-associated TNNT2 mutation.33

Mutations in genes for Z-disc elements in cardiomyopathy

Because mutations in the contractile elements were found in only less than half of HCM patients and only in a few DCM cases, there should be other disease gene(s) for HCM and DCM so that candidate gene approaches were taken to identify the disease-related mutations in other genes expressed in cardiac muscle (Figure 1). Identification of a HCM-associated mutation in titin gene (TTN) was the first example of disease gene other than the contractile elements,34 and the functional alteration due to the TTN mutation was an increased binding to α-actinin34. In addition, it was demonstrated that the HCM-associated Tcap gene (TCAP) mutations increased the binding of Tcap to titin, MLP and carsarcin-1,35 leading to a hypothesis that Z-disc mutations in HCM may result in increased binding of Z-disc components rendering sarcomere stiff. ‘Stiff sarcomere’ would increase passive tension upon stretch of sarcomere. Because the increased passive tension was associated with an increased Ca2+ sensitivity,36, 37, 38 it was speculated that HCM-associated abnormality in both Z-disc components and contractile elements might commonly result in the increased Ca2+ sensitivity of cardiac muscle. It should be noted that a possible controversy exists, that is, HCM-associated MLP gene (CSRP3) mutations were reported to decrease the binding to α-actinin and N-RAP.39, 40 However, DCM-associated mutations were also found in CSRP3 and α-actinin gene (ACTN2) and these mutations decreased binding to each other.41 Therefore, the decreased binding between MLP and α-actinin was associated with both HCM and DCM. This discrepancy should be resolved by further studies.

Figure 1
figure 1

Schematic representation of sarcomere components. Half sarcomere is schematically shown. Components are underlined when disease-associated mutations have been identified in patients with primary cardiomyopathy. A full color version of this figure is available at the Journal of Human Genetics journal online.

Full size image

Mutations in Z-disc components were relatively frequent in DCM patients (Table 3). We have reported several DCM-associated Z-disc component mutations in TTN,42, 43 CSRP3,44 TCAP35, 44 and Cypher/ZASP gene (LDB3).45 It was demonstrated that the DCM-associated TCAP mutations showed opposite functional alterations to the HCM-associated mutations.35 Similarly, a DCM-associated TTN mutation found in the actinin-binding domain showed a decreased binding to actinin.42 In addition, another DCM-associated TTN mutation found in the Tcap binding domain decreased the binding to Tcap.42 Because the Z-disc element mutations result in decreased binding among the elements, we hypothesize that DCM is the disease of ‘loose sarcomere’.12, 35 The loose sarcomere is evident in an animal model of DCM, a CSRP3 (MLP) knock-out mouse, in which Z-disc was wide and stretch response was impaired.44 Because the stretch response is a hypertrophic response of cardiomyocytes against passive tension and Z-disc elements have a role of stretch sensor in cardiomyocytes, abnormality in the Z-disc elements may alter the regulation of stretch response. In addition, molecular mechanisms for developing DCM due to the CSRP3 mutation44 were investigated in a knock-in mouse model.45

Cypher/ZASP is a Z-disc element connecting calsarcin and actinin.46 Calsarcin binds calcineurin,47 a Ser/Thr phosphatase involved in the process of hypertrophic program of cardiomyocytes.48 Functional significance of calcineurin anchorage to Z-disc is not fully understood, but it is involved in stress-induced calcineurin-NFAT activation, because heterozygous CSRP3 knock-out mice showed reduction in NFAT activation along with a dislocation of calcineurin from Z-disk.49 On the other hand, Cypher/ZASP binds protein kinase C46 and a DCM-associated LDB3 mutation in the protein kinase C-binding domain was found to increase the binding,50 suggesting that phosphorylation/dephosphorylation of Z-disc elements might be involved in the stretch response. Several other LDB3 mutations not in the protein kinase C-interacting domain were reported in DCM or LVNC.51 We searched for a binding protein to Cypher/ZASP by using the yeast two-hybrid method, and found phosphoglucomutase-1 (PGM1) as a novel binding protein.52 PGM1 is an enzyme involved in the glucose/glycogen metabolism and we demonstrated that PGM1 was localized at the Z-disc under the stressed culture conditions, low serum and low glucose, suggesting the role of PGM1 in the energy metabolism at the Z-disc.52 Because the DCM-associated mutations decreased the binding to PGM1,52 it was suggested that the decreased stress response might be involved in the pathogenesis of DCM.

There are several other DCM-associated mutations found in genes for the other Z-disk associated proteins, desmin (DES)53 and metavinculin (VCL).54 The VCL mutation impaired the binding to actin,54 while the DES mutations resulted in a subtle change in the cytoplasmic desmin network.55 In addition, mutations in myopalladin gene (MYPN) have recently been reported in DCM. Although the molecular mechanisms of MYPN mutations leading to DCM remained unclear, the DCM-associated mutations impaired the myofiblinogenesis.56

TTN mutations in cardiomyopathy

Titin is the most giant protein so far known and it spans from the Z-disc to M-line of sarcomere.57 As noted above, cardiomyopathy-associated mutations were found in the Z-disc region of titin,42 but a large body of DCM-associated mutations were found in other regions than the Z-disc region, that is, in the I-band region and M-band region. Initial reports of TTN mutations found in DCM patients included nonsense mutations and frameshift mutations42, 43 encoding truncated titin proteins. Recently, a systematic analysis of entire TTN coding exons has revealed that TTN is a major disease gene for DCM, because TTN mutations were found in about 26.7% of DCM patients; in addition to that; about 1.3% of HCM patients and 2.8% of healthy controls carried TTN mutations.58 In addition, it was found that most of the TTN mutations found in DCM patients were suggested to encode truncated proteins, while those in HCM were missense mutations.58 On the other hand, it was reported that TTN mutations could be found in many DCM families, but not all of them were co-segregated with the disease.59 In addition, TTN mutations including nonsense and frameshift mutations could also be found in healthy individuals, raising a possibility that not all of the TTN mutations found in patients with cardiomyopathy were disease-causing mutations.60 Recently, a large systematic analysis of TTN mutations has demonstrated that truncated mutations, that is, nonsense, frameshift and splicing site mutations, were found in both DCM patients and healthy controls, but the distribution of mutations was different between them. Mutations found in DCM patients were significantly enriched in the 3′ half of TTN, especially in the region encoding for A-band region.61 It is therefore suggested that the rare variants of TTN in the A-band region may be genetic risk factors for heart failure and it is not clear enough at the moment whether the TTN rare variants by themselves could cause DCM as the monogenic disease gene mutations with low penetrance or they are involved in the pathogenesis of DCM as genetic risk factors for polygenic disease. Nevertheless, TTN truncation mutations could contribute to development of heart failure, suggesting that the analysis of TTN would be useful in predicting clinical courses of DCM patients.61

Membranous and cytoskeletal mutations in cardiomyopathy

Identification of mutation in dystrophin gene (DMD) in male siblings of X-linked DCM62 was the first discovery of a disease gene for DCM. X-linked DCM is a rare form of familial DCM almost exclusively affecting males.63 DMD mutations are known to cause muscular dystrophy, which mainly affects skeletal muscles, and cardiac involvement is usually observed later in the clinical course.64, 65 In contrast, X-linked DCM cases usually manifest with cardiac symptoms and subtle skeletal muscle involvement,63 and phenotypic variance of DMD mutations may be explained by detecting which domain of dystrophin was affected.64 As shown in Table 3, DMD mutations could be found in 5% of Japanese sporadic DCM cases. None of these patients showed skeletal muscle symptoms, demonstrating that X-linked DCM should be considered not only for the male sibling of familial DCM but also for male cases of sporadic DCM.

Dystrophin is a membranous protein having a key role in mechanical links from extracellular matrix to intracellular cytoskeleton in association with other proteins forming a dystroglycan complex (DGC).66 Because muscle contraction forces deformity of myocytes with shortening/stretching, myofilaments should be tightly anchored to membrane and extracellular matrix via DGC to properly transmit the force with avoiding damages of cell membrane. Components of DGC in skeletal and cardiac muscles include dystrophin, α- and β-dystroglycans, α-laminins, α-, β-, γ- and δ-sarcoglycans, α- and β-dystrobrevins, syntrophin, and caveolin-3. In addition to DGC, α- and β-integrins are concentrated at the costameres that overly Z-lines in striated muscles, and the integrin complex also has a crucial role in mechanical links of power transmission.66 Therefore, abnormalities in DGC and integrin complex may result in muscular dystrophy and cardiomyopathy. Indeed, mutations in δ-sarcoglycan gene (SAGD),67 laminin α4 gene (LMNA4)68 and integrin-linked kinase gene (ILK)68 were found to cause DCM of autosomal dominant inheritance (Table 1).

Other mutations in cardiomyopathy

There are several other disease genes for HCM, including mutations in caveolin-3 gene (CAV3),69 meta-vinculin gene (VCL),70 αB-crystallin gene (CRYAB),71 junctophilin-2 gene (JPH-2),72 obscurin gene (OBSCN)73 and CARP gene (ANKRD1)74 (Figure 1). Functional analyses were reported for CRYAB, CAV3, OBSCN and ANKRD1 mutations; aggregation of αB-crystalline in cytoplasm,71 decreased cell surface expression of caveolin-3,69 decreased binding to titin73 and increased binding to titin and myopalladin,74 respectively. It is not clear how the aggregated αB-crystalline resulted in cardiac hypertrophy, but impaired stress response may exaggerate hypertrophic response.75 It is interesting to note that an HCM-associated TTN mutation in N2B region increased binding to FHL2 protein76 and decreased binding to αB-crystalline.77 Although function of obscurin is not fully understood, it may be involved in a calmodulin/CaMK-mediated signaling, because obscurin was reported to tether calmodulin to titin,78 which was impaired by the HCM-associated OBSCN mutation. The functional significance of increased binding of CARP to titin and myopalladin caused by the ANKRD1 mutations is not fully clarified, but mutant CARPs showed nuclear or perinuclear localization, whereas normal CARP was exclusively localized in the cytoplasm.79 Because CARP is a hypertrophy-related transcriptional co-factor79 and is known to be localized in the cytoplasm but shifted to nucleus upon stretching cardiomyocytes,80 abnormal nuclear localization of mutant CARPs suggested that the mutations rendered cardiomyocytes hypersensitive to the stretch response leading to hypertrophy. It was recently reported that three different ANKRD1 (CARP) mutations, all of which increased binding to titin/myopalladin and localized to nuclei, showed different functional changes in reduced stability of CARP proteins and affected contractility parameters of cardiac muscles in the engineered heart tissues,81 demonstrating that functional impacts of ANKRD1 mutations in cardiac hypertrophy might be quantitatively and qualitatively different in vivo.

On the other hand, there are many other disease genes for DCM, which may be categorized into several groups. The first group includes mutations in genes for nuclear membrane-associated proteins, lamin A/C (LMNA)82, 83 and emerin (EMD),84 which cause autosomal dominant and X-linked Emery-Dreifuss muscular dystrophy,85 respectively. Although molecular mechanisms underlying the development of DCM caused by the nuclear membrane abnormality remain not fully understood,86 a study of a LMNA mutation knock-in mouse87 showed that the mutation activated the MAPK pathway, suggesting an impaired signal transduction was involved in the pathogenesis of DCM.88 The second group includes mutations affecting ion channel function; regulatory subunit of ATP-sensitive potassium channel (ABCC9)89 and cardiac sodium channel (SCN5A).90 Clinical phenotypes of ABCC9 and SCN5A mutations were DCM accompanied by ventricular tachyrcardia89 and conduction defects,90 respectively. It should be noted here that the channelopathy is etiologically overlapping with the cardiomyopathy, such as SCN5A mutations in DCM and long QT syndrome, CAV3 mutations in HCM and long QT syndrome, and RYR2 mutations in ARVC and catecholaminergic polymorphic ventricular tachycardia.91 The third group is composed of mutations in genes for titin-N2B interacting proteins, four and half LIM protein (FHL2)92 and αB-crystallin (CRYAB).77 Because a titin-N2B region mutation found in DCM reduced binding to FHL276 and a FHL2 mutation reduced binding to titin-N2B,92 impaired interaction between titin and FHL2 was associated with DCM. Molecular mechanisms underlying this phenomenon may be that FHL2 function as a tethering molecule of adenylkinase, phosphofructokinase and muscle creatine kinase, that is, proper recruitment of metabolic enzymes was impaired, although abnormality in other functions of FHL293 could not be neglected. The DCM-associated CRYAB mutation decreased binding to titin-N2B region and a DCM-associated titin-N2B region mutation decreased binding to αB-crystallin,77 suggesting that impaired interaction between titin-N2B and αB-crystallin resulted in DCM. However, another HCM-associated titin-N2B mutation also reduced the binding to αB-crystallin.76 It is not clarified why the impaired binding of titin-N2B and αB-crystallin could exert both HCM and DCM phenotypes. There might be additional factors involved in the phenotypic expression of titin-N2B mutations, such that binding to FHL2 was different between the HCM- and DCM-associated mutations and that the DCM-associated titin-N2B mutation was a truncation mutation, whereas the HCM-associated mutation was a missense mutation.76 The fourth group is related to intracellular Ca2+ handling. Phospholamban is an inhibitory molecule of SERCA, which is activated when phosphorylated by protein kinase A.94 Functional analysis of phospholamban gene (PLN) mutations found in DCM showed that the mutations rendered phospholamban constitutive active, that is, inhibiting SERCA.95, 96 In contrast, a truncation mutation of PLN, that is, loss of PLN function, is reported in familial HCM.97 Although PLN deficiency in mice resulted in enhanced contractility,98 no cardiac hypertrophy was observed in the mice. In addition, loss of PLN rescued DCM phenotype99 in CSRP3 knockout mice, and a dominant-negative form of PLN prevented heart failure in cardiomyopathic hamster BIO14.6,100 which is known to be caused by SAGD deficiency.101 These observations suggest that functional impairment of phospholamban may prevent systolic dysfunction but not directly involved in the cardiac hypertrophy. Moreover, promoter mutations of PLN, which increased transcription, were recently reported in HCM.102, 103 Furthermore, we recently reported a FHOD3 mutation in DCM.104 FHOD3 is preferentially expressed in the heart and has a role in cellular actin polymerization in myofibrinogenesis and may be involved in repair of disrupted actin fibers. In cytoplasm, actin molecules can be found in both monomer and polymer forms, of which cellular stress induces actin polymerization; the process is called as actin dynamics. Monomeric actin is known to associate with a SRF coactivator MAL/MRTF-A/MKL1, and when actin polymerization would occur MAL/MRTF-A/MKL1 binds SRF to translocate into nucleus leading to induction of SRF-dependent gene expression.104 Functional assay showed that the FHOD3 mutation found in DCM impaired the actin dynamics-dependent SRF activation,104 suggesting that maladaptive cardiac remodeling would occur in the presence of FHOD3 mutation. In addition, mutations in Bcl2-associated athanogene 3 (BAG3) were reported as the cause of DCM.105, 106 Our functional analyses demonstrated that DCM-associated BAG3 mutations render cardiomyocytes susceptible to metabolic stress leading to apoptosis, which was not observed for another myofibrillar myopathy-associated BAG3 mutaion.106

There are further other mutations reported in DCM patients—G4.5 gene (tafazzin, TAZ, Barth’s syndrome),107 fukutin gene (FKTN),108 desmoplakin gene (DSP)109 and plakoglobin gene (JUP)110 mutations. These mutations, however, were found in ‘syndromic’ DCM that is accompanied by disorders and/or dysfunction in skeletal muscle, skin or hair. An example is that FKTN mutation was not found in pure DCM, but was found in skeletal myopathy accompanied by DCM and an early sign of FKTN mutation-associated DCM was hyper-CKemia.111

Genetic modifiers for cardiomyopathy

It is well known that carriers of gene mutations do not always develop disease even in the monogenic disease and this phenomenon is called as ‘penetrance’ of disease-causing mutation. Penetrance is defined by the frequency of subjects with disease manifestation among mutation carriers. It is well known for HCM that the penetrance could vary with age, gender, physiological performance or exercise depending on each mutation. In addition, both intra- and inter-familial phenotypic variation can be observed for each mutation and this variation is explained at least in part by the genetic background, genetic modifiers, which modify the expressivity of disease-causing mutations.112, 113, 114 Because the identification of such modifier genes is important and useful in predicting clinical course of each patient with specific mutation, various attempts have been performed to find out gene polymorphisms associated with disease progression since the early era soon after the identification of disease-causing mutations in HCM. Although there are several initial reports noting, for example, the association of gene polymorphisms in RAS cascade,112, 113, 115 such studies were not designed for comparison of polymorphisms in subjects between the population with and without specific mutation and hence it was difficult to assess the modifying effect of each polymorphism in modifying disease expression of a specific mutation. In this regard, it is needed to investigate the phenotypic expression as a quantitative trait to identify disease-modifying loci in the population with specific mutation. However, such a study will require a large cohort of patients with cardiomyopathy carrying the specific mutations and it is almost impossible to design a human study. Therefore, animal models carrying a specific mutation would be required and there are several reports addressing the issue of such attempts. It has been reported as early as in 2001 that cardiac hypertrophy in mice carrying the MYH7 Arg403Asn mutation were considerably different in mouse strains, although responsible loci was not identified.116 It has recently been reported from the analysis of transcriptome in mice with the MYH7 Arg403Asn mutation that a four-and-a-half LIM domain protein 1 gene (FHL1) has a key role in modifying the cardiac phenotype.117 It has also been speculated that FHL1 might be a genetic factor to determine the gender difference in the cardiac phenotype of HCM observed in this model mice.117 In addition, systematic search for genetic loci involved in the strain-specific cardiac phenotypes in calsequestrin gene (CSQ) overexpressing transgenic mice showing DCM phenotype has identified at least seven different quantitative trait loci,118, 119, 120 among which only one locus was clarified as cardiac troponin I-interacting kinase gene (TNNI3K) at the gene level.121

Another issue of modifier is the gender difference. It is well known that males are more susceptible to cardiomyopathy than females, because male to female ratios in patients with HCM and DCM are 2.3 and 2.6, respectively, and male patients suffer from more severe cardiac hypertrophy and faster disease progression in HCM122 and heart failure in DCM, especially in DCM caused by LMNA mutations,123 than female patients. Molecular basis of the gender difference in HCM has not been deciphered, but it was speculated in the mouse model with the MYH7 Arg403Asn mutation that FHL1 might have a role in the gender difference, because it is located on the X chromosome.119 As for the gender difference in the LMNA-linked DCM, we recently revealed that the male sex hormone was responsible for worse prognosis of DCM from the analysis of a mouse model, in which a LMNA mutation, H222P, was knocked-in. The model mice developed DCM with gender difference87 and it was revealed that androgen receptor (AR) was translocated into nuclei in the absence of testosterone, when cardiomyocytes carried the LMNA H222P mutation.124 In addition, the nuclear translocation of AR could be observed with another LMNA mutation, R225X, which was identified in a DCM family in which male patients developed DCM earlier than female patients. Because the nuclear translocation of AR was depending on FHL2 and associated with SRF, resulting in the activation of SRF-dependent transcription of cardiac remodeling associated genes, it was suggested that the gender difference in the LMNA-linked DCM was a consequence of abnormal regulation of remodeling genes in the presence of androgens.124 It should be noted here that an AR blocker, flutamide, could ameliorate DCM in the mouse model, anti-androgen therapy might prevent progression of heart failure in specific cases with LMNA-linked DCM, although the nuclear translocation of AR was not observed with another LMNA mutation, delK32, which did not show apparent gender difference in a knock-in model mice.124, 125

Mutations in other cardiomyopathies

Disease-causing gene mutations can also be identified in other cardiomyopathies. For example, mutations in sarcomere proteins were found in RCM (Table 1). It is interesting to note that MYH7, TNNT2 and TNNI3 mutations were associated with RCM, HCM and DCM. Molecular basis of the differences between RCM-associated mutations and HCM-associated mutations was that the RCM-associated mutations showed much greater Ca2+ sensitization than the HCM-associated mutations, as demonstrated for TNNT2126 and TNNI3127 mutations. In accordance with these findings, it was reported that restrictive phenotype (RCM-like HCM) was uncommon in HCM and may represent a poor prognosis form with severe diastolic dysfunction.128 On the other hand, the mechanistic difference between RCM-associated mutations and DCM-associated mutations is not elucidated, but a gene dose effect could be involved in the difference, because RCM-associated TNNI3 mutation was found in heterozygous state,129 whereas the DCM-associated TNNI3 mutation was found in homozygous state.130

LVNC is a form of cardiomyopathy in which ventricular trabeculations was poorly developed, and mutations in MYH7,131 CACT,132 DES,53 LMNA,133 TAZ,134 DTNA135 and LDB351 were reported in LVNC (Table 1). Molecular mechanisms of the mutations in causing LVNC are not elucidated. In a mouse model, deficiency of BMP10 resulted in the LVNC phenotype.136 BMP10 is a member of the TGFβ family, which is expressed mainly in the heart, and has a key role in morphogenesis of the heart.137 Therefore, LVNC might be a developmental error in the hearts carrying the mutations in components of sarcomere and/or sarcolemma.

Another primary cardiomyopathy AVRC has also been investigated for mutations138 (Table 1). Because the ARVC-associated mutations can be found in genes for plakoglobin (JUP),139 desmoplakin (DSP),140 plakophilin-2 (PKP2)141 and desmoglein (DSG3),142 they were considered to disrupt cell–cell contacts via desmosomes. RYR2 mutations were also reported in ARVC,143 linking cardiomyopathy to channelopathy. Promoter variant of TGFβ3 was also reported in ARVC,144 but its pathological significance remains to be resolved.

Conclusion remarks

In this review, gene mutations found in the hereditary cardiomyopathy are summarized. Each family or patient has usually only one disease-causing mutation, but the primary cardiomyopathy is both clinically and etiologically heterogeneous even in a specific clinical type. Because different causes result in the same phenotype, there might be several pathways in the pathogenesis of primary cardiomyopathy, such that abnormalities in the Ca2+ sensitivity, stretch response and metabolic stress response. Intervention of these common pathways will be a therapeutic or preventive strategy for hereditary cardiomyopathy caused by different mutations.