Center for Experimental Medicine
University Medical Center Hamburg-‐Eppendorf
Evaluation of two molecular-‐based therapies in a mouse model of
hypertrophic cardiomyopathy
by
Doreen Stimpel
DissertationDepartment of Chemistry
Faculty of Mathematics, Informatics and Natural Sciences University of Hamburg
for the degree of Doctor of Natural Sciences
1. Referee: Prof. Dr. rer. nat. Peter Heisig 2. Referee: Prof. Dr. med. Thomas Eschenhagen
Disputation: 19th December 2013
For my beloved family
T
ABLE OF CONTENTS
1 INTRODUCTION ... 1
1.1 Hypertrophic cardiomyopathy ... 3
1.2 The sarcomere – fundamental contractile unit ... 6
1.3 Cardiac myosin-‐binding protein C ... 8
1.3.1 Structural role ... 8
1.3.2 Functional role ... 10
1.3.3 MYBPC3 mutations and pathophysiological mechanisms ... 11
1.4 Current status of treatment and potential novel therapies for HCM ... 13
1.4.1 Drug and surgical treatments ... 14
1.4.2 Molecular-‐based treatments ... 16
1.4.2.1 Spliceosome-‐mediated RNA trans-‐splicing ... 16
1.4.2.2 Exon skipping/-‐inclusion ... 21
1.4.2.3 Conventional gene therapy ... 22
1.5 Gene delivery tools: AAV as a suitable vector for gene transfer ... 24
1.6 Aim of the work ... 28
2 MATERIALS AND METHODS ... 30
2.1 Materials ... 30
2.1.1 Mybpc3-‐targeted knock-‐in mouse model ... 30
2.1.2 Cell lines and isolated neonatal mouse cardiomyocytes ... 30
2.2 Methods ... 31
2.2.1 Design and cloning of PTMs ... 31
2.2.2 Design and cloning of ‘toxic’ molecules ... 35
2.2.3 Design and cloning of full-‐length WT-‐Mybpc3 ... 36
2.2.4 Restriction digestion ... 36
2.2.5 Ligation ... 37
2.2.6 Transformation of One Shot® TOP10 chemically competent E. coli ... 37
2.2.7 Preparation of plasmid DNA ... 37
2.2.8 Culture and transient transfection of HEK293 cells ... 38
2.2.9 Production of recombinant adeno-‐associated virus serotypes 6 and 9 ... 38
2.2.10 Production of recombinant adenovirus ... 39
2.2.11 Isolation and culture of cardiomyocytes ... 40
2.2.12 AAV6-‐mediated gene transfer in cardiomyocytes ... 40
2.2.13 Adenovirus-‐mediated gene transfer in cardiomyocytes ... 41
2.2.14 DNA and RNA analysis ... 41
2.2.14.1 RNA isolation ... 41
2.2.14.2 Determination of the RNA and DNA concentration ... 41
2.2.14.3 Reverse transcription ... 41
2.2.14.4 Polymerase chain reaction ... 42
2.2.14.5 Quantitative PCR ... 44
2.2.14.6 Agarose gel electrophoresis ... 46
2.2.14.7 Preparative agarose gels ... 46
2.2.15.2 Determination of protein concentration ... 47
2.2.15.3 Immunoprecipitation of proteins ... 47
2.2.15.4 Western blot ... 48
2.2.15.5 Immunofluorescence analysis of cardiomyocytes ... 49
2.2.16 In vivo studies ... 49
2.2.16.1 AAV9-‐mediated gene transfer in Mybpc3-‐targeted KI mice ... 49
2.2.16.2 In vivo bioluminescence imaging ... 49
2.2.16.3 Echocardiography ... 50
2.2.16.4 Organ extraction ... 50
2.2.16.5 Immunofluorescence analysis of cardiac sections ... 51
2.2.17 Statistical analysis ... 51
3 RESULTS ... 52
3.1 Mybpc3-‐targeted knock-‐in: a mouse model of HCM ... 52
3.2 The 5’-‐trans-‐splicing approach in Mybpc3-‐targeted KI mice ... 54
3.3 Characterization of the engineered constructs ... 56
3.3.1 Validation of PTMs ... 56
3.3.2 Validation of ‘toxic’ molecules ... 59
3.3.3 Validation of full-‐length WT-‐Mybpc3 ... 60
3.4 Evaluation of 5’-‐trans-‐splicing ex vivo ... 61
3.4.1 Validation of AAV6-‐mediated GFP expression in isolated cardiomyoctes ... 61
3.4.2 Characterization of 5’-‐trans-‐splicing at mRNA level ... 62
3.4.2.1 Detection of repaired Mybpc3 mRNA ... 63
3.4.2.2 Evaluation of 5’-‐trans-‐splicing in absence of the polyadenylation signal in the PTM ... 66
3.4.2.3 Validation of repaired Mybpc3 mRNA by sequencing ... 68
3.4.2.4 Semi-‐quantification of repaired Mybpc3 mRNA ... 69
3.4.3 Characterization of 5’-‐trans-‐splicing at protein level ... 71
3.4.3.1 Validation of potential translation of PTM and PTM∆pA in HEK293 cells ... 71
3.4.3.2 Detection of repaired cMyBP-‐C protein ... 72
3.4.4 Strategy to increase the efficiency of 5’-‐trans-‐splicing ... 77
3.4.5 Summary: 5’-‐trans-‐splicing partially corrected defective Mybpc3 mRNA ex vivo ... 80
3.5 Evaluation of 5’-‐trans-‐splicing in vivo ... 82
3.5.1 Validation of AAV9-‐mediated GFP expression in the heart ... 82
3.5.2 Evaluation of 5’-‐trans-‐splicing in adult Mybpc3-‐targeted KI mouse ... 83
3.5.2.1 Characterization of cardiac function ... 83
3.5.2.2 In vivo bioluminescence imaging ... 85
3.5.2.3 Characterization of 5’-‐trans-‐splicing at mRNA level ... 86
3.5.2.4 Characterization of 5’-‐trans-‐splicing at protein level ... 88
3.5.3 Evaluation of 5’-‐trans-‐splicing in neonatal Mybpc3-‐targeted KI mouse ... 89
3.5.3.1 Characterization of the cardiac function ... 90
3.5.3.2 Characterization of 5’-‐trans-‐splicing at the mRNA level ... 91
3.5.3.3 Characterization of 5’-‐trans-‐splicing at the protein level ... 93
3.5.4 Summary: 5’-‐trans-‐splicing partially corrected defective Mybpc3 mRNA in vivo ... 94
3.6 Conventional gene-‐based therapy in Mybpc3-‐targeted KI mice ... 95
3.6.1 Overexpression of WT-‐Mybpc3 in isolated cardiomyocytes ... 95
3.6.1.1 Characterization at mRNA level ... 95
3.6.1.2 Characterization at protein level ... 98
3.6.3 Summary: WT-‐Mybpc3 gene transfer rescued protein haploinsufficiency and prevented
accumulation of poison peptides in a mouse model of HCM ... 106
4 DISCUSSION ... 107
4.1 Mybpc3-‐targeted knock-‐in mice as a suitable model for molecular therapy ... 108
4.2 The combination of virus plus promoter plus delivery route for efficient cardiac gene transfer ... 110
4.2.1 Adeno-‐associated virus versus adenovirus ... 110
4.2.2 Choice of a suitable promoter ... 113
4.2.3 Vector delivery techniques in vivo ... 114
4.3 Molecular-‐based approaches to target the cause of HCM ... 115
4.3.1 RNA-‐based therapy to prevent the accumulation of poison polypeptides? ... 115
4.3.2 Conventional gene therapy to treat haploinsufficiency in HCM? ... 121
4.4 Conclusion -‐ future directions: from mice to men ... 124
5 SUMMARY ... 127
6 BIBLIOGRAPHY ... I
7 APPENDIX ... XV
7.1 Materials ... XV
7.1.1 Recombinant adeno-‐associated virus ... XV
7.1.2 Recombinant adenovirus ... XV
7.1.3 Antibodies ... XVI
7.1.3.1 Antibodies used for Western blot ... XVI
7.1.3.2 Antibodies used for immunofluorescence staining ... XVI
7.1.4 Bacterial strains ... XVII
7.1.5 Chemicals ... XVII
7.1.6 Consumable materials ... XIX
7.1.7 Kits ... XX
7.1.8 Laboratory equipment ... XX
7.1.9 Restriction enzymes ... XXI
7.1.10 Oligonucleotides ... XXII
7.1.10.1 Mybpc3 primer sequences for WT-‐Myppc3 ... XXII
7.1.10.2 Mybpc3 primer sequences used for PTM coding domain ... XXII
7.1.10.3 Mybpc3 primer sequences used for PTM binding domain ... XXII
7.1.10.4 Mybpc3 primer sequences for PCR and sequencing ... XXII
7.1.10.5 Mybpc3 primer and probe sequences for qPCR ... XXIII
7.1.10.6 Sequences of the binding domains ... XXIII
7.1.10.7 Sequences of the promoters ... XXIV
7.1.10.8 Vectors ... XXIV
7.2 H-‐ & P-‐Phrases ... XXV
7.3 Protein and DNA markers ... XXIX
9 DECLARATION ... XXXIV
10 CURRICULUM VITAE ... XXXVI
1 Introduction
"Cardiomyopathies are a heterogeneous group of diseases of the myocardium associated with mechanical and/or electrical dysfunction that usually (but not invariably) exhibit inappropriate ventricular hypertrophy or dilation and are due to a variety of causes that are frequently genetic. Cardiomyopathies either are confined to the heart or are part of generalized systemic disorders, often leading to cardiovascular death or progressive heart failure-‐related disability" (Maron et al., 2006).
Worldwide the estimated prevalence of all types of cardiomyopathies is about 3% in the general population (Cecchi et al., 2012). Cardiomyopathies, frequently with a genetic cause, include a variety of myocardial disorders in which the heart muscle is structurally and functionally abnormal (Elliott et al., 2008). The degree of cardiac dysfunction ranges from lifelong symptomless forms to major health problems, such as progressive heart failure, arrhythmia, thromboembolism or sudden death (Franz et al., 2001). A majority of patients remains undiagnosed or misdiagnosed with more prevalent cardiac conditions (Cecchi et al., 2012). Whereas major progress has been made in improving the prognosis of affected patients, cardiomyopathies still remain a considerable challenge in the health care system and an economic burden across Europe and the rest of the world.
Since the 1950’s several definitions, nomenclatures and classification schemes have been acquired by experts. Cardiomyopathies are currently grouped by the ‘European Society of Cardiology Working Group on Myocardial and Pericardial Diseases’ into specific morphological and functional characteristics with sub-‐classifications into familial and non-‐ familial subset (Figure 1; Elliott et al., 2008). The four major subtypes of cardiomyopathies are hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), arrhythmogenic right ventricular cardiomyopathy (ARVC) and restrictive cardiomyopathy (RCM). The classification also includes an unclassified group with no typical phenotype. The main characteristics of the different subtypes can be summarized as follows:
Hypertrophic cardiomyopathy is characterized by non-‐dilated ventricular chambers, but an
unexplained thickening of the left ventricle, due to an enlargement of cardiomyocytes and therefore the ventricle is less able to relax and to fill with blood.
Dilated cardiomyopathy is a weakness in the walls of the heart that causes dilation of the
left ventricle, compromising the heart’s efficiency and increasing the risk of congestive heart failure, arrhythmias and the formation of blood clots.
Arrhythmogenic right ventricular cardiomyopathy occurs if the muscle tissue in the right
ventricle is replaced with fibrofatty tissue, which disrupts the heart’s electrical signaling and causes ventricular arrhythmias.
Restrictive cardiomyopathy involves loss of elasticity of the ventricles due to stiff tissue that
prevents the ventricles from normal relaxation and from adequately blood filling prior contraction.
Figure 1: Classification scheme of cardiomyopathies -‐ European Society of Cardiology. HCM, hypertrophic
cardiomyopathy; DCM, dilated cardiomyopathy; ARVC, arrhythmogenic right ventricular cardiomyopathy; RCM, restrictive cardiomyopathy; idiopathic, no identifiable cause (Elliott et al., 2008).
This long-‐standing classification of cardiomyopathies is a multi-‐disciplinary approach to predict major complications, improve risk stratification and optimize treatment in each subtype. However, many cardiomyopathies are caused by a variety of gene abnormalities and are the consequence of interactions between multiple disease genes, unidentified
For example, different mutations within the same gene can result in different subtypes, e.g. mutations in MYH7 gene, encoding beta-‐myosin heavy chain, can cause either hypertrophic or dilated cardiomyopathy (Kamisago et al., 2000). In contrast, the same mutation in one distinct gene can arise with diverse cardiac phenotypes at different ages within the same family. The classification of cardiomyopathies based on morphological and functional phenotypes is an essential tool for clinicians to manage these complex heart diseases. The list of putative causative genes or non-‐genetic causes responsible for the distinct phenotype is provided for each subtype in the literature and is required to be constantly updated by diverse expert committees (Cecchi et al., 2012).
This work focuses on hypertrophic cardiomyopathy (HCM), the most common inherited cardiovascular disorder. It provides state of the art on HCM, including consequences of disease-‐causing mutations and describes how targeting the molecular defects have given early promise for potential new therapies.
1.1 Hypertrophic cardiomyopathy
Hypertrophic cardiomyopathy (HCM) is a common, but unexplained structural abnormality of the cardiac muscle and the most prevalent cause of heart-‐related sudden death in young people. The prominent morphologic feature of HCM is a massive asymmetrical left ventricular hypertrophy (LVH), which mainly involves the interventricular septum (Figure 2). Clinically diagnosed HCM patients show increased left ventricular wall thicknesses ranging from mild (15 mm) to massive (˃30 mm) magnitudes based on echocardiographic measurements (Maron, 2002). The abnormal thickening of the left ventricle may result in left ventricular outflow tract obstruction and is associated with an initially normal systolic function, whereas the diastolic function is impaired (Maron, 2002). In fact, disease progression may lead to a thinning of the left ventricle wall and an enlargement of the left cavity, which is associated with reduced ejection fraction. This significantly increases the risk of irreversible heart failure and unexpected sudden cardiac death, especially in young individuals during exercise (Maron et al., 2000). HCM occurs mostly in the absence of any other cardiac or systemic disorders that themselves would be able to trigger hypertrophy in
remarkable variability in disease development, age of onset and severity of symptoms, showing both benign and malignant manifestations (Richard et al., 2003). Indeed, many patients display a clinical history completely asymptomatic, whereas in other cases HCM leads to symptoms such as vertigo, chest pain, syncope, dyspnea and can turn into malignant arrhythmias and progressive heart failure (Gersh et al., 2011). The main histological features include chaotically oriented cardiomyocytes (=myocardial disarray), as well as myocyte hypertrophy and increased interstitial fibrosis (Figure 2).
Figure 2: Histological characteristics of hypertrophic cardiomyopathy. Upper scheme shows hypertrophied and
normal hearts. The lower part represents corresponding histological section stained with hematoxylin and eosin. Hypertrophied cardiomyocytes display disarray and increased myocardial fibrosis. Figure adapted from the Mayo clinic website (upper panel) and from Ho et al., 2010 (lower panel).
HCM has a prevalence of 1:500 in the general population and occurs equally in both sexes (Maron et al., 1995). It accounts for 36% of sudden cardiac death among competitive athletes (Maron et al., 1996). In the majority of cases HCM is inherited in an autosomal-‐ dominant pattern (Richard et al., 2003). The variability of the HCM phenotype is attributed to over 1000 mutations in at least 19 different genes (Table 1), which have been identified as a potential cause of the disease (Schlossarek et al., 2011, Friedrich and Carrier, 2012). Most known genes encode components of the contractile apparatus, the sarcomere, which elucidates HCM as a ‘sarcomeropathy’. Most of the HCM patients are heterozygous for the mutation and 3-‐5% of them carry two independent mutations at once resulting in a more severe phenotype than patients with a single mutation (Richard et al., 2003, Ingles et al.,
mutation (Ho et al., 2000). Mutations in the MYH7 gene encoding cardiac beta-‐myosin heavy chain and in the MYBPC3 gene encoding cardiac myosin-‐binding protein C are the most common ones and comprise more than 70% of the known HCM-‐causing genetic defects (Richard et al., 2006, Maron et al., 2012). The phenotypic expression of MYBPC3 mutations is largely heterogeneous. In contrast to most MYH7 mutations, which cause early onset and extensive LVH (Watkins et al., 1992), mutations in MYBPC3 have been first shown to be associated with delayed onset, incomplete penetrance and mild hypertrophy (Charron et al., 1998, Niimura et al., 1998). However, severe cases with a poor outcome and high sudden cardiac death risk profile have been reported as well (Erdmann et al., 2001, Oliva-‐Sandoval et al., 2010). Especially childhood HCM is often associated with an extreme LVH and the presence of sinus and supraventricular tachycardia (Morita et al., 2008, El-‐Saiedi et al., 2013). Besides the heterogeneity of the causal genes and mutations, there is a high variability in the phenotype expression of HCM (Marian, 2002). Even within single families, affected individuals with identical causal mutations can show significant variability in the disease penetrance, age of onset and clinical manifestation. In fact, 20-‐30% of HCM mutation-‐positive patients do not reveal any cardiac phenotype (Richard et al., 2003). This suggests that phenotypes are not exclusively gene or mutation specific and distinct modifiers must exist, such as environmental factors, epigenetic signaling, microRNAs, gene polymorphisms or posttranslational modifications, which modulate disease-‐causing mechanisms (Marian, 2002, Richard et al., 2006, Schlossarek et al., 2011). For example, it has been reported that a polymorphism in the promoter of the CALM3 gene, encoding calmodulin III, has modifying impact on the HCM phenotype by affecting the expression level of CALM3 and consequently the calcium handling and development of LVH (Friedrich et al., 2009). Other potential modifier genes are polymorphisms in genes encoding angiotensin I-‐ converting enzyme (Tesson et al., 1997), as well as the AT1 and AT2 receptors (Osterop et al., 1998, Deinum et al., 2001)
Table 1: Summary of sarcomeric gene mutations implicated in HCM (Schlossarek et al., 2011).
1.2 The sarcomere – fundamental contractile unit
The basic functional and structural unit of contractile muscles is the sarcomere with a highly ordered assembly. The cardiac sarcomere is formed by three types of myofilaments, thin and thick filaments and titin (Figure 3). Their organization within the sarcomere is defined by two neighboring Z-‐lines, which enclose the I-‐Band (thin filaments) and the A-‐band (thick filaments) with the M-‐line and C-‐zone. The thin filaments contain α-‐cardiac actin, α-‐ tropomyosin and the troponin complex. The latter is composed of troponin C (binds calcium), troponin I (inhibits contraction) and troponin T (binds to α-‐tropomyosin). Actin filaments interact with α-‐actinin in the Z-‐line and extent to the A-‐band. The thick filaments consist of mainly myosin and are located in the A-‐band. Myosin consists of two heavy chains (α/ß myosin heavy chain, MHC) and four light chains (MLC; two essential light chains and two regulatory light chains). The third filament titin stabilizes the thick filaments and connects them to the Z-‐line.
Figure 3: Scheme of the cardiac sarcomere representing the localization of the cMyBP-‐C protein. The thin
filaments are composed of α-‐cardiac actin, α-‐tropomyosin and the troponin complex. Thick filaments are located in the A-‐band and consist of myosin with α/ß myosin heavy chains, essential and regulatory myosin light chains. cMyBP-‐C is a thick filament-‐associated protein, aside titin, which itself is considered as the elastic component of the sarcomere. The I-‐band, A-‐band, with the C-‐zones and M-‐line are confined by two Z-‐lines (Schlossarek et al., 2011).
The interaction of actin and myosin is responsible for the muscle contraction and is triggered by an increase in cytosolic calcium through voltage-‐dependent L-‐type calcium channels during action potential. This inward calcium flux activates the ryanodine receptors in the sarcoplasmic reticulum and induces calcium release. In the resting cardiomyocyte, α-‐ tropomyosin blocks the myosin-‐binding site on actin and is anchored by troponin T and I. When intracellular calcium binds to troponin C, it initiates a conformational change in the troponin complex and releases α-‐tropomyosin from the myosin-‐binding site on actin, allowing the interaction of myosin and actin. Upon adenosine triphosphate hydrolysis, myosin is subjected to a series of conformational changes resulting in the motion of the thick along the thin filaments. Several calcium pumps, such as sarcoplasmic reticulum calcium-‐ ATPase (SERCA) and sodium-‐calcium exchanger contribute to the removal of calcium to return to the relaxed state. Finally, α-‐tropomyosin is again locked in its actin-‐blocking position by troponin T and I.
1.3 Cardiac myosin-‐binding protein C
1.3.1 Structural roleThe cardiac myosin-‐binding protein C (cMyBP-‐C) is a multidomain protein of the thick filaments and is located in doublets in the C-‐zone of the A-‐band of the sarcomere (Figure 3). cMyBP-‐C protein bundles the thick filaments transversally over nine clear stripes in each half A-‐band (Luther et al., 2008). The MYBPC3 gene encoding the cMyBP-‐C protein, is located on the human chromosome 11p11.2 (Gautel et al., 1995). Its complete structure and sequence was established in 1997 (Carrier et al., 1997). The gene comprises more than 21,000 bp and contains 35 exons, of which 34 are coding. The transcript is translated into a 150-‐kDa protein, which exhibits fundamental structural and regulatory functions (Winegrad, 1999, Flashman et al., 2004, de Tombe, 2006, Granzier and Campbell, 2006, Schlossarek et al., 2011, Sadayappan and Tombe, 2012).
Beside the cardiac isoform of MyBP-‐C there are two others, which have been identified in adult cross-‐striated muscle, each encoded by different genes: the slow-‐skeletal (MYBPC1) and the fast-‐skeletal (MYBPC2) isoforms. All three isoforms belong to the intracellular immunoglobulin superfamily and show a conserved pattern in their main structure. Generally, MyBP-‐C protein consists of seven immunoglobulin (Ig-‐1-‐like) and three fibronectin (FN-‐3) domains (C1-‐C10). The cardiac isoform is exclusively expressed in the mammalian heart (Gautel et al., 1995, Fougerousse et al., 1998) and differs from the other isoforms in distinct characteristics. It contains an extra Ig-‐1-‐like domain (C0) located at the N-‐terminus, four phosphorylation sites located in the MyBP-‐C motif, which is a conserved 105-‐residues linker between C1 and C2, a 28-‐amino acid insertion in the C5 domain and a proline-‐alanine-‐ rich extension between C0 and C1 (Figure 4; Gautel et al., 1995, Flashman et al., 2004, Oakley et al., 2004).
Figure 4: Schematic representation of MYBPC3 gene, MYBPC3 mRNA and structure of cMyBP-‐C protein. In the
upper part the organization of the MYBPC3 gene 5’ to 3’ is shown with localization of exons indicated by boxes. In the middle panel the mRNA of the joined exons after cis-‐splicing is displayed. cMyBP-‐C protein domains involved in sarcomeric protein interactions are indicated by arrows. Abbreviations: P, phosphorylation site; S2, myosin subfragment S2; LMM, light meromyosin (Schlossarek et al., 2011).
The cardiac isoform of MyBP-‐C protein interacts with different sarcomeric proteins via specific motifs or domains (Figure 4). The MyBP-‐C motif is known to interact with the subfragment S2 of myosin (Gruen and Gautel, 1999), whereas the C10 domain binds to light meromyosin (Okagaki et al., 1993) and the C8-‐C10 domains to titin (Freiburg and Gautel, 1996). Potential actin-‐binding sites were described at the proline-‐alanine-‐rich extension between C0 and C1 domains (Squire et al., 2003), at the C0 domain (Kulikovskaya et al., 2003) and at the C1-‐C2 domains (Razumova et al., 2006). Interactions of cMyBP-‐C protein with myosin and titin are important for an optimal arrangement of the sarcomere and the complex binding enables to form a very stable structure (Flashman et al., 2004). The precise cMyBP-‐C protein incorporation into the thick filament setup still remains unclear, but the ‘trimeric collar’ model is preferred. This model suggests the trimerization of three cMyBP-‐C molecules in a staggered parallel orientation around the backbone of the thick filaments
(Moolman-‐Smook et al., 2002). This ‘trimeric collar’ formation leads to interactions between C5 and C8, C6 and C9, C7 and C10 of two cMyBP-‐C molecules and C10 of the third one binds to the myosin rod (Figure 5). The distinct C0-‐C4 domains interact with the myosin subfragment S2 and actin thin filaments.
1.3.2 Functional role
cMyBP-‐C protein regulates the cross-‐bridge cycling via titin, myosin and actin interactions, the myofilament calcium sensitivity and relaxation of the sarcomere. Its regulatory role is mediated through four phosphorylation sites located in the MyBP-‐C motif (Barefield and Sadayappan, 2010). Phosphorylation occurs in response to beta-‐adrenergic stimulation by cAMP-‐dependent protein kinase (PKA) (Gautel et al., 1995), by calcium/calmodulin-‐ dependent kinase II (CaMKII) in a calcium-‐dependent manner (McClellan et al., 2001), by protein kinase C ε (PKCε) (Kooij et al., 2010), by protein kinase D (PKD) (Bardswell et al., 2010) and by 90-‐kDa ribosomal S6 kinase (RSK) (Cuello et al., 2011). Upon phosphorylation of cMyBP-‐C protein, the binding to actin and myosin subfragment S2 is abolished, which increases the cross-‐bridging between myosin and actin and therefore the force of contraction (Figure 5) (Flashman et al., 2004). It is therefore essential for a normal cardiac function and was shown to be cardioprotective (Sadayappan et al., 2005, Sadayappan et al., 2006). It has been reported that the level of phosphorylated cMyBP-‐C protein is low in human and experimental models of heart failure (El-‐Armouche et al., 2007) as well as in myocardial tissue of HCM patients (van Dijk et al., 2009, Marston et al., 2012, van Dijk et al., 2012). Moreover, cMyBP-‐C protein is crucial to allow a complete diastolic relaxation of the sarcomere at low intracellular calcium concentrations by inhibiting the interaction of actin and myosin through reversible binding to the myosin subfragment S2 (Kulikovskaya et al., 2003, Pohlmann et al., 2007). Residual cross-‐bridge cycling in diastole, incomplete relaxation and increased calcium sensitivity of the myofilaments may lead to diastolic dysfunction, hypercontractility and increased energy usage (Crilley et al., 2003, Javadpour et al., 2003, Keller et al., 2004, Pohlmann et al., 2007). The detailed role of cMyBP-‐C protein is still not fully understood, but alteration in protein level and phosphorylation may lead to severe cardiac dysfunction and structural abnormalities, in particular to cardiac hypertrophy.
Figure 5: Sarcomeric organization of cMyBP-‐C protein in the dephosphorylated and phosphorylated state.
Three cMyBP-‐C molecules trimerize around the thick filament backbone of light-‐meromyosin (Myosin LMM) and titin. The N-‐terminal C1-‐M-‐C2 domains in the dephosphorylated state are tightly attached to myosin-‐S2 and actin via reversible linkage to prevent the cross-‐bridge formation. Upon phosphorylation, the motif (M domain) releases the interaction with myosin-‐S2 and actin and results in an attachment of the myosin heads (myosin-‐S1) to the thin filaments, which promotes strong actin-‐myosin interaction (Schlossarek et al., 2011).
1.3.3 MYBPC3 mutations and pathophysiological mechanisms
More than 460 different mutations associated with HCM have been found in the MYBPC3 gene (source HGMD; http://www.hgmd.org/). However, the expression of the mutations and therefore the consequences at mRNA and protein levels are not known for most of them. The majority (˃70%) of MYBPC3 mutations are frameshift or nonsense and predicted to cause altered splicing (Richard et al., 2006, Carrier et al., 2010, Marian, 2010). Nonsense mutations directly introduce a premature termination codon (PTC) in the transcribed mRNA, whereas frameshift mutations are the consequence of point mutations, insertions or deletions, which lead to a PTC downstream of the mutation in the transcript. C-‐terminal truncated cMyBP-‐C proteins, likely lacking myosin-‐ and/or titin-‐binding sites (Carrier et al., 1997, Richard et al., 2006) have never been detected in myocardial tissue of patients carrying MYBPC3 mutations at the heterozygous state (Rottbauer et al., 1997, Moolman et al., 2000, Marston et al., 2009, van Dijk et al., 2009, Marston et al., 2012, van Dijk et al., 2012). The full-‐length cMyBP-‐C protein level in myectomy samples from HCM patients with frameshift MYBPC3 mutations is 20-‐30% lower than in normal hearts (Marston et al., 2009, van Dijk et al., 2009, Marston, 2011). Apparently, the gene product from the remaining
functional wild-‐type allele cannot fully compensate for the defect transcript from the mutant allele.
Therefore, the molecular mechanisms of MYBPC3-‐associated HCM mutations and their impact at mRNA and protein levels are not completely conclusive. The reduced amount of normal cMyBP-‐C protein is one argument that haploinsufficiency is likely the HCM disease mechanism. The expression of the functional wild-‐type allele in the case of heterozygous mutation does not produce a sufficient amount of protein to maintain the wild-‐type phenotype. Haploinsufficiency has been also reported in several animal models of MYBPC3-‐ associated HCM. The Mybpc3-‐tageted KI mouse model express low levels of cMyBP-‐C protein, 80% in the heterozygous state and only 10% in the homozygous state (Vignier et al., 2009). The Maine Coon cats, carrying a natural MYBPC3 missense mutation reveal 69% lower level of cMyBP-‐C protein in the heterozygous state and 88% in the homozygous state than in wild-‐type controls (Meurs et al., 2005). Since sarcomere stoichiometry is tightly regulated, reduced levels of cMyBP-‐C protein could imbalance the assembly of the thick filament and therefore affect sarcomeric structure and function, leading to contractile deficits. Furthermore, the cMyBP-‐C protein deficiency is likely associated with higher myofilament calcium sensitivity due to the altered protein expression and/or its phosphorylation, which
represents a consistent abnormality in HCM (Harris et al., 2002, Cazorla et al., 2006, van Dijk et al., 2009). The myofilament activation results in basal cardiomyocyte hypercontractility and excessive energy usage (Watkins et al., 2011). The resultant energy deficiency and altered intracellular calcium handling combined with activation of signaling pathways likely contribute to the anatomic (hypertrophy, myocardial disarray and fibrosis) and functional (diastolic dysfunction) characteristics of HCM (Ashrafian et al., 2011).
Additionally, the presence of truncated mutant cMyBP-‐C proteins could abnormally alter the sarcomere organization and potentially provoke damage in cardiomyocytes acting as poison peptides. However, despite detectable amounts of mutant mRNA (25-‐45% of wild-‐type), truncated proteins have not been detected in myocardial tissue of HCM patients (Moolman et al., 2000, Marston et al., 2009, van Dijk et al., 2009, Marston et al., 2012, van Dijk et al., 2012). This suggests that their expression is regulated at mRNA and protein levels. In the cell the major quality control systems to regulate expression of nonsense and frameshift
(UPS) and autophagy-‐lysosomal pathway (ALP) (Sarikas et al., 2005, Carrier et al., 2010, Schlossarek et al., 2011). The NMD degrades nonsense transcripts at mRNA level and UPS and/or the ALP rapidly eliminate misfolded or mutant proteins (Vignier et al., 2009). The permanent degradation of mutant proteins by the UPS and/or ALP to protect the cell from their deleterious effects may lead to an impairment of the proteolytic control systems. In heterozygous Mybpc3-‐targeted KI and KO mouse models, it has been reported that adrenergic stress or aging resulted in saturation of the systems (Schlossarek et al., 2012a, Schlossarek et al., 2012b). The potential accumulation of poison polypeptides was sufficient to alter cell homeostasis and trigger the disease progression (Sarikas et al., 2005, Bahrudin et al., 2008). Impairment of the UPS has been reported in human HCM patients as well (Predmore et al., 2010).
1.4 Current status of treatment and potential novel therapies for
HCM
As mentioned before, HCM is a very complex disease with heterogeneous genetic, morphologic, functional and clinical manifestation. Although it is a life-‐threatening disease, no curative treatment exists up to date reversing the cardiac hypertrophy and dysfunction and/or preventing sudden cardiac death (Carrier et al., 2010, Schlossarek et al., 2011, Frey et al., 2012, Spoladore et al., 2012). The standard clinical management is basically empiric. Current drug-‐based strategies are partially capable to relieve the HCM-‐associated symptoms and slow down the disease progression but none of them induces regression of cardiac hypertrophy or fibrosis or targets the genetic cause. Therefore, potential innovative therapies against fundamental pathophysiological mechanisms in patients with inherited HCM mutations urgently need to be established and may display a new paradigm for personalized medicine.
1.4.1 Drug and surgical treatments
Current drug-‐based interventions of HCM mainly focus on symptomatic management and on the control of ventricular outflow obstruction and arrhythmias in order to improve the patient’s quality of life. Although clinical guidelines for diagnosis and treatment of HCM exist (Maron et al., 2003, Gersh et al., 2011), the benefit of pharmacological intervention is not evidence-‐based (Spoladore et al., 2012). Until 2012, only less than 50 pharmacological studies have been performed, most of them involving small and non-‐randomized groups of patients and none of them has prospectively addressed the long-‐term outcome (Figure 6) (Spoladore et al., 2012). Long-‐standing drugs, such as beta-‐adrenergic inhibitors and L-‐type calcium channel blockers have been applied in the majority of studies.
Figure 6: Number of HCM pharmacological studies (left) and number of patients included (right) based on the
application of the indicated drugs. ACE-‐I, angiotensin converting enzyme inhibitors; ARBs, angiotensin receptor blockers (Spoladore et al., 2012).
The beneficial effect of beta-‐blockers, such as propranolol, is due to improvement of ventricular relaxation, increase in time for diastolic filling and reduction of excitability, especially in patients with exercise-‐induced symptoms, left ventricular outflow obstruction and chest pain (Marian, 2009). The negative inotropic response of beta-‐blocker additionally decreases the myocardial oxygen demand and outflow gradient during exercise. L-‐type calcium channel blocker, such as verapamil and diltiazem are employed to lower the heart rate, reduce excitability and lengthen the diastolic filling period in non-‐obstructive HCM (Spirito et al., 1997, Marian, 2009). Besides the first line therapy of cardiac arrhythmias with beta-‐blocker, amiodarone is commonly used for the treatment of atrial and ventricular arrhythmias in HCM patients (McKenna et al., 1985, Cecchi et al., 1998). Disopyramide, a
class I antiarrhythmic drug with a negative inotropic effect, has been successfully used in combination with beta-‐blocker to attenuate symptoms in patients with left ventricular outflow obstruction (Sherrid et al., 2005). Very limited data exist for drugs targeting the altered energy homeostasis (perhexeline) (Abozguia et al., 2010), the impaired calcium cycling and sensitivity of the myofilaments (blebbistatin) (Baudenbacher et al., 2008), the increased fibrosis and left ventricular remodeling (aldosterone antagonists, such as spironolactone or angiotensin II receptor blockers, such as losartan, irbesartan) (Lim et al., 2001, de Resende et al., 2006). If symptoms persist despite drug-‐based treatment, surgical septal myectomy or percutaneous septal ablation with ethanol could be employed in order to mechanically diminish the outflow tract obstruction, to attenuate the severity of symptoms and reduce the risk of sudden cardiac death (Maron, 2002, Elliott and McKenna, 2004, Ball et al., 2011). The insertion of an implantable cardioverter-‐defibrillator (ICD) as prophylactic intervention may be indicated in individuals who have survived cardiac arrest or those with an increased susceptibility for atrial and ventricular arrhythmias (Ho, 2010). ICD has been reported to be effective and life-‐saving in relevant patients (Begley et al., 2003, Maron et al., 2007). Patients with end-‐stage HCM, which is characterized by left ventricular remodeling with progressive wall thinning, cavity enlargement and systolic dysfunction, should be medicated with appropriate drugs for heart failure including diuretics and angiotensin converting enzyme inhibitors (ACE-‐I) (Spirito et al., 1987, Spirito et al., 1997). Patients with severe heart failure ultimately require heart transplantation (Shirani et al., 1993).
1.4.2 Molecular-‐based treatments
The concept of genetic medicine in order to repair or modify inherited disorders was initially evoked in the late 1960s with the development of virus-‐based transformation of mammalian cells and the progress in recombinant DNA techniques (Friedmann, 1992, Sheridan, 2011). Molecular-‐based therapy involves the use of DNA or RNA for the treatment, curing or prevention of disorders. It generally aims at the correction of key pathologies, which are out of reach for conventional drugs. Depending on the nature of the disease, gene-‐based approaches can be applied to deliver a functional, therapeutic gene to substitute the defective or missing endogenous gene analogue (conventional gene therapy) or to reduce the level of defective transcripts using innovative RNA-‐based approaches, such as spliceosome-‐mediated RNA trans-‐splicing (SMaRT), exon skipping and exon inclusion. The idea is to utilize site-‐directed gene or RNA editing strategies to target critical molecular changes in the endogenous gene or pre-‐mRNA to anticipate for causal HCM therapy. Mutations, causing aberrant splicing represent approximately about one third of all disease-‐ causing mutations (Lim et al., 2011, Sterne-‐Weiler et al., 2011). The splicing mechanism, as an early step in gene expression, is an attractive intervention point for therapeutic purposes, which does not alter the genome. During the last 10 years approaches targeting mutant pre-‐ mRNA in a splice-‐switching manner have been intensively studied in the field of neuromuscular genetic disorders (Le Roy et al., 2009, Havens et al., 2013). These studies used sophisticated molecular tools, including pre-‐trans-‐splicing molecules and modified antisense oligonucleotides.
1.4.2.1 Spliceosome-‐mediated RNA trans-‐splicing
Spliceosome-‐mediated RNA trans-‐splicing (SMaRT) is a promising therapeutic approach for genetic disorders. It is a post-‐transcriptional process occurring during mRNA maturation, and is defined as a splicing reaction between two independently transcribed RNA molecules, a target endogenous mutant pre-‐mRNA and a therapeutic pre-‐trans-‐splicing molecule (PTM) (Wally et al., 2012). Trans-‐splicing is a rare process, but naturally occurring mechanism and was first discovered in lower eukaryotes, such as trypanosomes (Sutton and Boothroyd,
important process to achieve functional diversity (Dorn and Krauss, 2003) and has been recently reported in human leukocytes (Chiu et al., 2008). Trans-‐splicing is very attractive since it occurs by using the endogenous spliceosome machinery in the nucleus, it does not require the introduction of the complete gene and it is restricted to those cells expressing the target pre-‐mRNA. Briefly, eukaryotic post-‐transcriptional processing of pre-‐mRNA includes 5’-‐capping, 3’-‐polyadenylation and cis-‐splicing reactions, which occur in the nucleus. It has been intensively studied that addition of 7-‐methylguanosine to the 5’-‐end and 3’-‐ polyadenylation of the transcript are essential for proper splicing, RNA transport and protection from degradation. Additionally, these processes enhance translation of mRNA (Sachs and Wahle, 1993, Wahle and Keller, 1996, Colgan and Manley, 1997, Cowling, 2010). In general, cis-‐splicing is a multistep process to remove introns from the pre-‐mRNA and ligate remaining exons to form a single continuous mRNA molecule. The cis-‐splicing reaction is catalyzed by a protein complex called the spliceosome, consisting of several proteins and small nuclear RNA molecules that recognize splice sites within the pre-‐mRNA sequence. These universally conserved sequences in eukaryotic pre-‐mRNA are the 5’-‐ or donor-‐splice site (GU) and 3’-‐ or acceptor-‐splice site (AG) at exon-‐intron boundaries, the conserved branch point (A) and a pyrimidine-‐rich region (Py)n just upstream of the 3'-‐splice site (Figure
7; Pagani and Baralle, 2004).
Figure 7: Classical cis-‐splicing reaction and essential intronic splicing signals. The image shows conserved
nucleotides at the exon-‐intron boundaries: 5’-‐splice site (GU), branch point (A), polypyrimidine tract (Py)n and
3’-‐splice site (AG). Splicing involves two-‐step transesterification reactions. In the first step the 2’-‐hydroxyl-‐ group of a specific branch point nucleotide within the intron performs a nucleophilic attack on the phosphate (p) of the 5'-‐splice site forming the lariat structure. The second transesterification reaction involves the 3’-‐ hydroxyl group of the released 5’-‐exon and the phosphate (p) at the 3’-‐splice site, which releases the lariat structure and ligates the two exons (Pagani and Baralle, 2004).
SMaRT is an emerging technology carried out by the endogenous spliceosome complex. The engineered PTM is exogenously delivered to the nucleus and after successful transcription it hybridizes with the target pre-‐mRNA to finally generate a chimeric mRNA molecule (Figure 8). The PTM is designed to recode a specific part of the mRNA by suppressing cis-‐splicing while enhancing trans-‐splicing in a competitive manner; therefore, as the level of repaired mRNA increases the level of native mRNA should decrease. This observation suggests SMaRT as a general approach for correction of a targeted pre-‐mRNA transcript (Puttaraju et al., 1999).
Figure 8: Schematic illustration of SMaRT strategy. Pre-‐trans-‐splicing molecule (PTM) is dispensed to the cell
and transcribed in the nucleus. The transcript targets the pre-‐mRNA of the gene of interest (GOI) and produces repaired mRNA and protein. This process is in competition with cis-‐splicing, which gives rise to the endogenous mRNA and proteins (Mearini et al., 2013).
PTMs consist of three main domains: i) a coding domain containing the wild-‐type exonic information; ii) a set of splice signals containing 5´-‐ and/or 3´-‐splice sites, including the branch point sequence; iii) a binding domain complementary to intronic sequences of the target pre-‐mRNA by base pairing. The latter is essential to put the splicing sites of both pre-‐ mRNA and PTM close to each other in order to induce trans-‐splicing. The length of the binding domain defines the specificity of the PTM and has a profound impact on the trans-‐ splicing activity (Mansfield et al., 2004). Usually the size of the binding domain comprises 70-‐ 150 nucleotides (Puttaraju et al., 1999), although no formula for its length has been defined