Hypertrophic cardiomyopathy (HCM) is the most common cardiac genetic disease, with more than 1400 different mutations in genes encoding primarily sarcomeric proteins (Friedrich and Carrier, 2012; Maron et al., 2014; Ho et al., 2015). The most frequently mutated genes are MYH7 (encoding β-myosin-heavy chain) and MYBPC3 (encoding cardiac myosin-binding protein C), which constitute about 80% of known mutations. Besides a typical hypertrophy of the left ventricle, patients often present a normal or increased ejection fraction, but a compromised diastolic function with an incomplete relaxation and increased filling pressures (Elliott et al., 2014). Diastolic dysfunction may result in left atrial enlargement and is associated with exercise intolerance and bad prognosis in HCM, primarily due to supraventricular arrhythmias (Yang et al., 2009). Tissue Doppler measurements have revealed that a reduction in systolic and diastolic velocities is prominent even before the development of left ventricular hypertrophy (Charron et al., 1997).

Increased myofilament Ca²⁺ sensitivity, as observed in three Mybpc3 cardiomyopathy mouse models (Mybpc3 KO and KI) developed by us and others (Cazorla et al., 2006; Pohlmann et al., 2007; Vignier et al., 2009; Fraysse et al., 2012; Barefield et al., 2014), and in other animal models of HCM (Knollmann et al., 2001; Robinson et al., 2007; Iorga et al., 2008), could be an underlying cause of diastolic dysfunction. This observation has also been made in human HCM samples (Jacques et al., 2008; van Dijk et al., 2009, 2012) and could explain the incomplete relaxation in diastole in MYBPC3-associated HCM (and probably other cases associated with an increased Ca²⁺ sensitivity). Additionally, myofilaments with increased sensitivity to Ca²⁺ may act as Ca²⁺ buffers, prolonging the export of Ca²⁺ and relaxation time which could contribute to diastolic dysfunction and arrhythmias (Morimoto et al., 1998; Baudenbacher et al., 2008). These findings support the hypothesis that interventions decreasing myofilament Ca²⁺ sensitivity could reverse the phenotype of HCM and have therapeutic value (Jagatheesan et al., 2007; Alves et al., 2014; Tardiff et al., 2015).

Epigallocatechin-3-gallate (EGCg), a major component of green tea, has been suggested to be effective against cardiovascular diseases. Proposed mechanisms were anti-oxidative, anti-inflammatory, vasorelaxant, and positive inotropic effects (Chyu et al., 2004; Lorenz et al., 2004; Ludwig et al., 2004). Furthermore, it was shown that EGCg lowered myofilament Ca²⁺ sensitivity in a transgenic HCM mouse model expressing a human cardiac troponin T (TNNT2, cTnT) mutant (Tadano et al., 2010) and in HCM-associated human cardiac troponin I (TNNI3, cTnI) and cTnT mutants in a reconstituted acto-myosin system (Warren et al., 2015; Messer et al., 2016). However, the effects of EGCg were not evaluated in other HCM models associated with mutations in the thick filament of the sarcomere. Since MYBPC3 is the major disease gene constituting 45% of genetically diagnosed HCM cases (Ho et al., 2015), we used a representative mouse model carrying the human c.772G> A MYBPC3 mutation (Vignier et al., 2009). This mutation was found in 14% unrelated HCM patients in Tuscany and is associated with a bad prognosis (Richard et al., 2003; Girolami et al., 2006; Ho et al., 2015). These mice exhibit, in addition to left ventricular hypertrophy and decreased fractional area shortening, increased myofilament Ca²⁺ sensitivity, and diastolic dysfunction (Fraysse et al., 2012). We evaluated the acute effects of EGCg on sarcomere shortening and Ca²⁺ transient in intact ventricular myocytes and on force-Ca²⁺ relationship of skinned cardiac muscle strips isolated from KI and wild-type (WT) mice.

One major feature of HCM is a compromised diastolic function with an incomplete relaxation (Elliott et al., 2014). Increased myofilament Ca²⁺ sensitivity as seen in HCM patients and in several mouse models of HCM (Morimoto et al., 1998; Robinson et al., 2007; Huke and Knollmann, 2010; Kimura, 2010; Fraysse et al., 2012; Moore et al., 2012; van Dijk et al., 2012; Barefield et al., 2014; Flenner et al., 2016) could contribute to diastolic dysfunction. Recent findings advocate a potential therapeutic role for EGCg in HCM since it lowered the increased myofilament Ca²⁺ sensitivity in a transgenic HCM mouse model expressing a human cardiac troponin T (TNNT2, cTnT) mutant (Tadano et al., 2010) and in HCM-associated human cardiac troponin I (TNNI3, cTnI) and cTnT mutants in a reconstituted acto-myosin system (Warren et al., 2015; Messer et al., 2016). The aim of this study was to evaluate whether EGCg has beneficial effects in another thick filament model of HCM, carrying a human mutation in the thick filament protein gene MYBPC3. The main findings of this study were: 1. At baseline and in the absence of Fura-2, KI cardiomyocytes exhibited higher fractional sarcomere shortening, lower diastolic sarcomere length, longer contraction and relaxation times than WT cells, without differences in Ca²⁺ transient amplitude and kinetics. 2. EGCg had no effect on sarcomere shortening or diastolic sarcomere length, but it accelerated relaxation and Ca²⁺ transient decay in Mybpc3 WT and KI cardiomyocytes. 3. EGCg induced myofilament Ca²⁺ desensitization in permeabilized left ventricular trabeculae isolated from Mybpc3 WT and KI mouse hearts. 4. EGCg effects on relaxation time and myofilament Ca²⁺ sensitivity were more pronounced in KI cells and muscle strips, respectively.

EGCg is a major component of green tea and has been reported to have beneficial effects in a variety of diseases (Peng et al., 2011; Brückner et al., 2012; Ortsäter et al., 2012). Suggested mechanisms in the context of cardiovascular diseases are anti-oxidative, anti-inflammatory, vasorelaxant and positive inotropic effects (Chyu et al., 2004; Lorenz et al., 2004; Ludwig et al., 2004). Since HCM patients typically present with a normal or even increased ejection fraction, but with a diminished diastolic function (Elliott et al., 2014), mimicked in Mybpc3 KI mice (Fraysse et al., 2012), we intended to apply the highest concentration that would not increase sarcomere shortening (1.8 μM) to evaluate its effects on the impaired relaxation in KI myocytes. Similar to Lorenz et al. and Tadano et al. (Lorenz et al., 2008; Tadano et al., 2010), who reported a positive inotropic effect between 2.5 and 5 μM in mouse hearts and rat myocytes, we observed an increase in fractional sarcomere shortening at concentrations above 3 μM. Although we observed no effect on diastolic sarcomere length in WT cells at 1.8 μM EGCg, we speculated that this concentration could have an effect in KI cells. This was not the case. Nevertheless, 1.8 μM EGCg reduced relaxation time, and this effect was more prominent in KI cells, whereas the effects on diastolic Ca²⁺ and Ca²⁺ kinetics did not differ between the genotypes. The increase in diastolic Ca²⁺ could be explained by an inhibiting effect on the Na⁺/Ca²⁺ exchanger (NCX), as reported with 10 nM (Feng et al., 2012) and 2.5 μM (Lorenz et al., 2008). The faster time to peak Ca²⁺ is probably due to EGCg effects on the ryanodine receptor. Indeed, it has been shown that EGCg activates the ryanodine receptor at 10 nM in sarcoplasmic reticulum (SR) vesicles isolated from rabbit left ventricles (Feng et al., 2012) and in the range of 1 nM–20 μM in junctional SR vesicles isolated from skeletal muscle (Najafi et al., 2015). At a concentration of 1.8 μM, EGCg had no effect on PLB Ser16/Thr17 phosphorylation (Supplemental Figure 1). This supports findings of Lorenz et al., who neither observed an effect on PLB Ser16/Thr17 phosphorylation with EGCg (4 μM) nor an influence on EGCg actions on contractility after β1-adrenoceptor inhibition with 3 μM metoprolol (Lorenz et al., 2008). We thus exclude that the acceleration of relaxation and Ca²⁺ kinetics are mediated via the β1-adrenergic pathway. In contrast to Lorenz et al. we did not observe an increased Ca²⁺ peak height in the presence of 1.8 μM EGCg (Lorenz et al., 2008). We therefore assume that at this concentration EGCg does not increase sarcoplasmic reticulum (SR) Ca²⁺ load. Data from canine SR vesicles and HEK293 cell microsomes show that EGCg concentrations > 4.8 μM directly inhibit SR calcium ATPase (Kargacin et al., 2011). Additionally, Feng et al. reported no effect on SERCA in rabbit cardiac and skeletal SR membranes with 1–2 μM EGCg (Feng et al., 2012; Najafi et al., 2015). We therefore rule out that the acceleration of relaxation kinetics in the presence of 1.8 μM EGCg is due to an increased SERCA activity. A plausible contributing reason could be an EGCg-mediated decrease in myofilament Ca²⁺ sensitivity.

Indeed, it has been reported that EGCg lowered the increased myofilament Ca²⁺ sensitivity and improved the diastolic function of isolated working heart preparations from a transgenic HCM mouse model expressing a human TNNT2 mutation (Tadano et al., 2010). EGCg also restored the increased Ca²⁺ sensitivity of HCM-associated human cTnI and cTnT mutants in a reconstituted acto-myosin system (Warren et al., 2015; Messer et al., 2016). Similar to these previous data, 30 μM EGCg decreased Ca²⁺ sensitivity in our thick myofilament mouse model that carries a frequent HCM mutation in the most frequently mutated gene (Vignier et al., 2009; Fraysse et al., 2012; Ho et al., 2015). It has been suggested that EGCg binding to the C-terminal region of cardiac troponin C (cTnC) alters the interaction between cTnC and cTnI and therefore the sensitivity of the myofilaments to Ca²⁺ (Liou et al., 2008; Robertson et al., 2009). In both intact myocytes and skinned trabeculae, EGCg had a more profound effect on cells and strips of the KI genotype. This could also be related to the longer baseline relaxation time and elevated myofilament Ca²⁺ sensitivity in KI mice. This is similar to a recent study in which we showed that the myofilament Ca²⁺-desensitizing effect of ranolazine was only present in KI, but not in WT muscle strips (Flenner et al., 2016). Since EGCg is not a pure Ca²⁺ desensitizer (Stangl et al., 2007), side effects such as arrhythmias or blood pressure lowering reported in in vivo applications would also be expected in WT mice (Alvarez et al., 2006; Bao et al., 2015).

The study has two limitations. (1) The Mybpc3 KI model shows many HCM features only at the homozygous state. Additionally, Mybpc3 KI mice present a reduced ejection fraction. These two points are in contrast to the more common findings in HCM patients who present left ventricular hypertrophy, interstitial fibrosis, and diastolic dysfunction with heterozygous mutation states and normal or even supra-normal ejection fraction. (2) The Ca²⁺-desensitizing effect of EGCg on the permeabilized trabeculae was only detected at 30 μM, which is higher than the concentrations used in the intact myocyte experiments. This discrepancy has also been observed before (Tadano et al., 2010; Robinson et al., 2016). It should be kept in mind that in permeabilized cardiac muscle fibers the effective concentrations of the Ca²⁺-sensitizers pimobendan and EMD57033 were also reported to be much higher than those estimated in vivo (Fujino et al., 1988; Solaro et al., 1993; Chu et al., 1999), proposing that drugs could have lower potency in permeabilized cardiac muscle preparations than in vivo, since important components enhancing drug uptake or function such as membrane transporters or the SR could be disturbed in their function or missing after muscle strip skinning.

EGCg accelerated relaxation and Ca²⁺ transient decay in Mybpc3 WT and KI cardiomyocytes, which seems to be partly due to Ca²⁺ desensitization of the myofilaments. We show for the first time that EGCg is also effective in a thick filament mutation mouse model. In support of other studies (Tadano et al., 2010; Warren et al., 2015; Messer et al., 2016), this confirms that EGCg belongs to a new class of Ca²⁺ antagonists which have a very favorable functional profile acting directly on the Ca²⁺ regulatory system of cTnC. Further efforts should be made to develop a new generation of Ca²⁺ desensitizers based on the green tea catechin EGCg with even more potency and specificity to avoid potential side effects such as arrhythmias or lowering of blood pressure that could also affect healthy individuals.

FWF: conception and design of research, management of the mouse cohorts, execution of experiments, analysis, and interpretation of data, figure preparation, drafting of the manuscript. FF: isolation and treatment of cardiac myocytes, execution of experiments, interpretation of data, figure preparation, discussion of the manuscript draft. MN: isolation and treatment of cardiac myocytes, execution of experiments. TE: interpretation of data, discussion of the manuscript draft. LC: conception and design of research, analysis, and interpretation of data, drafting of the manuscript. All authors critically discussed the results, and reviewed and approved the manuscript before submission.

This work was supported by the DZHK (German Centre for Cardiovascular Research) and the Deutsche Stiftung für Herzforschung (F/28/12).