Ranolazine Prevents Phenotype Development in a Mouse Model of Hypertrophic Cardiomyopathy

Abstract

Background—

Current therapies are ineffective in preventing the development of cardiac phenotype in young carriers of mutations associated with hypertrophic cardiomyopathy (HCM). Ranolazine, a late Na⁺ current blocker, reduced the electromechanical dysfunction of human HCM myocardium in vitro.

Methods and Results—

To test whether long-term treatment prevents cardiomyopathy in vivo, transgenic mice harboring the R92Q troponin-T mutation and wild-type littermates received an oral lifelong treatment with ranolazine and were compared with age-matched vehicle-treated animals. In 12-months-old male R92Q mice, ranolazine at therapeutic plasma concentrations prevented the development of HCM-related cardiac phenotype, including thickening of the interventricular septum, left ventricular volume reduction, left ventricular hypercontractility, diastolic dysfunction, left-atrial enlargement and left ventricular fibrosis, as evaluated in vivo using echocardiography and magnetic resonance. Left ventricular cardiomyocytes from vehicle-treated R92Q mice showed marked excitation–contraction coupling abnormalities, including increased diastolic [Ca²⁺] and Ca²⁺ waves, whereas cells from treated mutants were undistinguishable from those from wild-type mice. Intact trabeculae from vehicle-treated mutants displayed inotropic insufficiency, increased diastolic tension, and premature contractions; ranolazine treatment counteracted the development of myocardial mechanical abnormalities. In mutant myocytes, ranolazine inhibited the enhanced late Na⁺ current and reduced intracellular [Na⁺] and diastolic [Ca²⁺], ultimately preventing the pathological increase of calmodulin kinase activity in treated mice.

Conclusions—

Owing to the sustained reduction of intracellular Ca²⁺ and calmodulin kinase activity, ranolazine prevented the development of morphological and functional cardiac phenotype in mice carrying a clinically relevant HCM-related mutation. Pharmacological inhibitors of late Na⁺ current are promising candidates for an early preventive therapy in young phenotype-negative subjects carrying high-risk HCM-related mutations.

Hypertrophic cardiomyopathy (HCM), despite being the most common inherited cardiac disease, currently does not have a disease-modifying therapeutic option.¹ Beta blockers, Ca²⁺ channel blockers, and disopyramide are commonly used to reduce symptoms associated with left ventricular (LV) outflow tract obstruction, diastolic dysfunction, and arrhythmias in HCM.² However, their long-term effectiveness in slowing disease progression and improving patient outcomes remains unclear and is likely to be marginal.³

Extensive family screening programs are identifying increasing numbers of young individuals who carry HCM-associated mutations with no or minimal phenotype and normal cardiac function.² Although genotype–phenotype correlations in individual patients with HCM remain unreliable in predicting outcome, growing evidence suggests that certain genes or variants may be associated with adverse outcome and higher likelihood of disease progression.⁴ A case in point is represented by patients with thin-filament mutations, who show an increased prevalence of sudden cardiac death in the pediatric age range⁵ and, later in life, a higher risk of progression toward end-stage heart failure,⁶ when compared with the more common thick-filament mutation carriers.

To date, however, when a young individual is diagnosed with having a high-risk HCM-associated mutation, clinicians can only suggest a close follow-up strategy, for lack of robust therapeutic options capable of delaying or preventing disease onset and progression.¹ Although gene therapy still seems distant and likely reserved to selected genotypes, an attractive alternative is represented by disease-modifying drugs interfering with the subcellular and cellular mechanisms of myocardial adverse remodeling.⁷ A growing number of preclinical studies have identified changes in intracellular Ca²⁺ handling as fundamental determinants of electromechanical dysfunction in HCM myocardium, both in transgenic animal models^8,9 and in human ventricular cardiomyocytes.¹⁰ Indeed, elevated intracellular Ca²⁺ levels are associated with increased activation of cardiac Ca²⁺–calmodulin-dependent protein kinase (CaMKII), a key player of adverse cellular remodeling in human HCM.¹⁰ HCM-related abnormalities of Ca²⁺ homeostasis seem to occur independently of the underlying mutation, thus, representing an ideal universal target.

Of note, previous studies pointed out that long-term reduction of intracellular Ca²⁺ overload with the Ca²⁺ channel blocker diltiazem is capable of delaying and mitigating phenotype expression in HCM mouse models.¹¹ Following these observations, a pilot clinical study has been performed in which nonhypertrophic mutation carriers exhibiting initial diastolic abnormalities were treated with diltiazem. The results of the study revealed that early treatment with diltiazem does delay the development of diastolic dysfunction in carriers.¹² However, the study duration was too short to measure a significant effect on the development of hypertrophy.¹² Furthermore, diltiazem is not the ideal drug in young asymptomatic patients because of potential occurrence of bradycardia, hypotension, and exercise intolerance.¹² Nevertheless, these observations are consistent with the idea that an early and sustained reduction in intracellular Ca²⁺ may slow the natural progression of disease phenotype.

An alternative therapeutic option to reduce intracellular Ca²⁺ overload in HCM is pharmacological inhibition of the late Na⁺ current (I_NaL) using the clinically available, prototypical inhibitor ranolazine. I_NaL is pathologically increased in human HCM, causing intracellular Ca²⁺ accumulation.¹⁰ Studies performed on human HCM cardiomyocytes showed that after I_NaL inhibition, reduction of action potential (AP) duration (APD) leads to decreased Ca²⁺ entry, while normalization of intracellular Na⁺ potentiates Ca²⁺ extrusion from the cytosol via the Na–Ca exchanger (NCX),¹⁰ ultimately causing a reduction of both [Ca²⁺] and CaMKII activity.¹³

Based on these findings, we sought to determine whether I_NaL inhibition with ranolazine affects long-term disease progression in HCM. Toward this goal, transgenic mice carrying the HCM-related R92Q mutation in the troponin-T (TnT) gene¹⁴ were subjected to a lifelong oral treatment with ranolazine and compared with vehicle-fed animals. Results of the present study showed that long-term treatment with ranolazine counters the development of HCM-related cardiac remodeling and prevents myocardial dysfunction in mutant mice.

Materials and Methods

Detailed methods are available in the Data Supplement. All experimental protocols were performed in accordance with current Italian and European regulations and were approved by the local institutional review board and by the animal welfare committee of the Italian Ministry of Health. We used male C57BL/6N knock-in mice carrying the R92Q mutation in the TnT gene (R92Q mice),^14,15 as well as wild-type (WT) littermates. Each litter was randomized to treatment with chow containing 0.5% ranolazine and 0.03% ketoconazole (RAN group) or ketoconazole only (KET group), starting 1 to 2 days after birth. Treatment was continued till 11 to 12 months of age when mice were used for the following experiments. Ranolazine plasma concentration at the time of killing was measured with liquid chromatography analysis coupled with tandem mass spectrometry. Echocardiography was performed on anesthetized mice as previously described¹⁶ to characterize LV morphology and systolic and diastolic function. Cardiac magnetic resonance was performed as previously described¹⁷ to asses LV and right ventricular volumes and mass, as well as to calculate the fraction of extracellular space by studying the decay of contrast signal after intravenous gadolinium compound infusion.¹⁸ Single cardiomyocytes were isolated from excised hearts via enzymatic dissociation and used for intracellular Ca²⁺ measurements using Ca-sensitive fluorescent dyes^19,20 to evaluate the amplitude and kinetics of Ca²⁺ transients, diastolic [Ca²⁺], and the rate of arrhythmic spontaneous Ca²⁺ release during stimulation with field electrodes. In a subset of experiments, simultaneous detection of intracellular [Na⁺] and [Ca²⁺] was performed after staining single cells with Fura-2 and Natrium Green-2. Patch-clamp studies were performed to record APs and I_NaL in single LV cardiomyocytes. To assess T-tubule density, myocytes were stained with membrane-selective dyes and observed with a confocal microscope.¹⁹ Formalin-fixed LV slices were stained with picrosirius red and used to assess intramyocardial fibrosis. Fast-frozen LV myocardial samples were processed to obtain total proteins, which were used for Western blot studies to assess the expression and phosphorylation of CaMKII and other proteins involved in excitation–contraction coupling.²¹ LV and right ventricular intact trabeculae were dissected from explanted hearts^20,22 and used to record isometric force during electric stimulation with different pacing protocols, aimed at evaluating twitch amplitude and kinetics, response to high-frequency pacing, prolonged pauses, as well as the effects of β-adrenergic activation.¹⁹ Skinned trabeculae were used to obtain force–pCa curves by transferring them manually between baths containing different pCa solutions, as previously described.²³ Data from cells and muscles are expressed as mean±SEM (number of samples and animals are indicated in the respective figure legends). Statistical analysis was performed as previously described²¹ using SPSS 23.0 (IBM) and STATA 12.0 (StataCorp). In brief, all sets of variables were checked for normality (Shapiro–Wilk test) and for homogeneity of variances among groups (Levene’s Test). The statistical tests used to calculate P values for each data set are indicated with abbreviations in the respective figure legends. For variables where a single measurement for each mouse is included (eg, echocardiography, Western blot), the 3 different groups were compared using (1) 1-way analysis of variance with Tukey correction (for normally distributed homoscedastic data sets), (2) Kruskal–Wallis test with Dunn’s multiple comparison test (for non-Gaussian data sets) or (3) Welch’s analysis of variance with Games–Howell test (for heteroscedastic groups). For variables where measurements from an unequal number of different samples (eg, cells or trabeculae) from each mouse are included (path-clamp, ion fluorescence, isometric force data), we used linear mixed models to compare data groups to account for intrasubject correlation; when comparing >2 groups, the Tukey–Kramer post hoc method was used to compute P values for all pairwise comparisons. P<0.05 was considered statistically significant.

Results

Rationale for Using Ranolazine in R92Q Mice

We first sought to determine whether and how acute ranolazine administration exerts beneficial effects in the myocardium of R92Q mutant mice. LV cardiomyocytes were isolated from the hearts of R92Q and WT male mice aged 4 to 6 months and used for patch-clamp experiments to record I_NaL and APs in the absence and presence of 10 μM ranolazine. I_NaL density was 2-fold higher in cardiomyocytes from R92Q as compared with those from WT cells. In mutant cardiomyocytes, ranolazine reduced I_NaL down to the level of that in WT cells (Figure 1A and 1B). Accordingly, ranolazine administration led to a reduction of APD in R92Q cardiomyocytes at all investigated frequencies, while it did not affect APD in WT cells (Figure 1C and 1D). Interestingly, I_NaL was already increased in cardiomyocytes from R92Q mice at 1 month of age, as compared with cells from age-matched WT mice (Figure IA in the Data Supplement), suggesting that the increase of I_NaL occurs early during disease development in this model. In line with previous reports,²⁴ ranolazine (10 μM) shows use-dependent block of I, with no significant effects on peak Na⁺ NaL current (Figure IB and IC in the Data Supplement).

Figure 1.

Acute effects of ranolazine in cardiomyocytes from R92Q-TnT mice. A, Representative superimposed traces of late Na⁺ current from wild-type (WT; left) and R92Q cardiomyocytes (right) in the absence and presence of 10 μM ranolazine. B, Average I_NaL density integrals, calculated from 25 to 125 ms after the onset of a 500 ms depolarization step from −120 to −10 mV. Means±SEM from 14 WT (3 mice) and 15 R92Q (3 mice) cardiomyocytes. C, Representative superimposed action potential traces from WT (left) and R92Q cardiomyocytes (right) elicited at 1 Hz in the absence and presence of 10 μM ranolazine. D, Average action potential duration at 90% of repolarization (APD90% at 1 Hz stimulation rate). Means±SEM from 21 WT cells (6 mice) and 17 R92Q cells (5 mice). E, Representative simultaneous fluorescence recordings of intracellular [Ca²⁺] (Fura-2, top) and [Na⁺] (Natrium Green-2, bottom) from a WT cell (left), a R92Q cardiomyocyte (center), and the same diseased cell in the presence of 10 μM ranolazine (right). Cells were field stimulated at 1, 3, 5, and 7 Hz for >45 s at each frequency. F, Average intracellular diastolic [Ca²⁺] at the different frequencies of stimulation. G, Average intracellular diastolic [Na⁺]. F and G, Means±SEM from 36 WT cells (3 mice) and 27 R92Q cardiomyocytes (3 mice). A–G, Black, WT cells at baseline; gray, WT with ranolazine; red, R92Q cells at baseline; violet, R92Q cells with ranolazine. *P<0.05, **P<0.01, with linear mixed model (LMM). #P<0.05, ##P<0.01, with LMM for paired measurements.

We then studied the effects of I_NaL inhibition on intracellular [Na⁺] and [Ca²⁺] by simultaneously monitoring the concentration of the 2 ions in R92Q and WT cardiomyocytes costained with the ion-sensitive fluorescent dyes Asante Natrium Green and Fura-2 (for Na⁺ and Ca²⁺, respectively). Cardiomyocytes were field-stimulated at increasing rates (1–7 Hz) in the absence and presence of ranolazine. Both intracellular [Na⁺] and [Ca²⁺] were higher in R92Q myocytes as compared with those in WT cells. In agreement with the observed I_NaL inhibition, ranolazine significantly reduced diastolic [Na⁺] and [Ca²⁺] in mutant cells, especially at high frequencies of stimulation (Figure 1E and 1F), while the amplitude of Ca²⁺ transients was slightly increased (Figure ID in the Data Supplement). Contrarily, the intracellular concentration of the 2 ions was only mildly affected by the drug in WT cells (Figure 1E and 1F; Figure ID in the Data Supplement). Interestingly, only untreated R92Q cardiomyocytes show a linear correlation between intracellular [Na⁺] and [Ca²⁺] (Figure IE in the Data Supplement); ranolazine abolishes the correlation, suggesting that I_NaL inhibition is able to fully counter the pathological increase of intracellular [Ca²⁺] that depends on Na⁺ overload. The remaining excess of [Ca²⁺] in R92Q cells after I_NaL inhibition likely depends on Na⁺-unrelated mechanisms and is unaffected by the drug at this stage.

Of note, all the observed effects of ranolazine on I_NaL, intracellular [Ca²⁺], and [Na⁺] were lost after 2 minutes of drug washout (Figure IF through IH in the Data Supplement).

Long-Term Treatment Protocol

To test whether ranolazine is effective in slowing or preventing the development of the pathological phenotype in mutant mice, male animals carrying the R92Q TnT mutation and WT littermates received an oral treatment with ranolazine since birth and were compared with age-matched vehicle-treated siblings. A detailed scheme of the experimental protocol is shown in Figure II in the Data Supplement. Male R92Q animals were fed with chow containing 0.5% ranolazine+0.03% ketoconazole (RAN treatment). Ketoconazole was added to inhibit hepatic CYP3A4 to maintain ranolazine plasma concentration within the range of 2 to 8 μmol/L throughout the day, values comparable to therapeutic levels in humans (Figure III in the Data Supplement). RAN-treated animals were compared with age-matched R92Q animals fed with chow containing only 0.03% ketoconazole (KET treatment). R92Q heterozygous mice were mated with WT females to obtain mixed litters. Treatment was started immediately after birth; each litter was assigned randomly to RAN or KET treatment group. Survival analysis did not show any difference among WT-KET, WT-RAN, R92QKET, and R92Q-RAN groups: the number of animal who died during the 12 months of treatment was low and comparable in the 4 groups (WT-KET=1/23, WT-RAN=1/22; R92Q-KET=2/22; R92Q-RAN=2/23; P>0.05 at χ² test for all comparisons). No differences were noted between WTKET and WT-RAN mice in terms of gross cardiac structure and function as evaluated with a specific subset of in vivo assessments (Table I in the Data Supplement). Therefore, WT-RAN mice were used as a single control group throughout the study (henceforth named WT). All animals were treated for 12 months and then used in the experiments described below. Average ranolazine concentration measured in plasma at killing was 4.41±1.01 μmol/L (mean±SE from 15 animals).

Echocardiography

Before echo assessments, heart rate and arterial blood pressure were measured in conscious mice: no differences were noted in the basal vital parameters among the 3 groups of animals (Table II in the Data Supplement). Echocardiographic acquisitions were performed in anesthetized 12-month-old male mice from the 3 study groups (11 WT, 11 R92Q-KET, and 11 R92Q-RAN) using a standardized protocol²⁵; measurements were performed offline in a blind manner.

Lifelong ranolazine treatment fully prevented the development of structural LV changes in R92Q mutation carrier mice (Figure 2). The increase in interventricular septum thickness, a hallmark of HCM remodeling, was not present in ranolazine-treated mice, as evaluated from long-axis parasternal views (Figure 2A and 2C). Three short-axis views (Figure 2B) at different levels of the LV were used to estimate chamber volumes. A reduction in end-diastolic and end-systolic LV volumes, although present in all vehicle-treated mutants, was entirely prevented in R92Q-RAN mice (Figure 2D). LV hypercontractility was documented in R92Q-KET animals by an increased ejection fraction (LV ejection fraction) and an augmented fractional increase of septal thickness during contraction, as compared with WT. Remarkably, LV hypercontractility was absent in ranolazine-treated mice (Figure 2E). No differences were noted in terms of stroke volume, cardiac output, and heart rate among the 3 groups of animals (Table III in the Data Supplement).

Figure 2.

Echocardiographic measurements. A, Representative parasternal long-axis views of the left ventricle at end diastole from wild-type (WT), R92Q-KET, and R92Q-RAN mice. Vertical bars mark the thickness of the interventricular septum. The length of horizontal bars equal 1 mm. B, Representative short-axis views (mid-left ventricle) at end diastole (left) and end systole (right) from mice belonging to the 3 study groups. Horizontal bars equal 1 mm. C, Thickness of the interventricular septum (IVS) measured at end diastole (left) and at end systole (right) in WT, R92Q-KET, and R92Q-RAN mice. D, Left ventricular volumes calculated using the Simpson technique at end diastole (EDV; left) and end systole (ESV; right) in mice from the 3 study groups. E, left ventricular ejection fraction (LV-EF; left) and IVS systolic thickening (right), expressed as percentage of the diastolic value, measured in mice from the 3 cohorts. C–E, Means±SEM from 11 mice per group. *P<0.05; **P<0.01; NS, P>0.05; statistical test: 1-way analysis of variance with Tukey correction. R92Q-KET indicates R92Q mutant mice fed with chow containing ketoconazole only; and R92Q-RAN, R92Q mutant mice fed with chow containing ranolazine and ketoconazole.

Doppler studies of transmitral blood flow (Figure 3A) and measurements of left atrial (LA) dimensions (Figure 3C) were performed in 4-chamber views to assess diastolic function. Reduced transmitral blood flow during early diastole (ie, reduced E wave amplitude) paralleled by an increased proportion of LV diastolic filling relying on atrial contraction (ie, increased A wave amplitude) is found to be consistent with impaired ventricular relaxation. LA size may increase because of elevated LV filling pressure. E/A ratio was markedly lower, and LA size increased in R92Q-KET mice compared with WT mice (Figure 3B and 3D). Ranolazine treatment partially prevented the impairment of diastolic function: in R92Q-RAN mice, LA dimensions were fully normalized (Figure 3D), and E/A ratio was significantly higher than that in R92QKET mice, though it was still lower than that in WT mice (Figure 3B).

Figure 3.

Diastolic function. A, Representative transmitral blood flow velocity curves recorded using pulsed-wave Doppler echocardiography in apical view from wild-type (WT), R92Q-KET, and R92Q-RAN mice. Vertical bar equals 100 mm/s. Early (passive, E) and late (atrial contraction, A) left ventricular (LV) filling waves are indicated on the traces. B, Ratio between A and E LV filling waves (E/A ratio) in mice from the 3 study cohorts. C, Representative 4-chamber apical views from WT, R92Q-KET, and R92Q-RAN mice. Left atrium area of each image is highlighted with a dotted line. Horizontal bar equals 1 mm. D, Left atrium areas in mice from the 3 study groups. B and D, Means±SEM from 11 mice per group. *P<0.05; **P<0.01; NS, P>0.05. Statistical tests: Welch’s analysis of variance in B, 1-way analysis of variance with Tukey correction in C. LA indicates left atrium; LAA, left atrial appendage; LV, left ventricle; RA, right atrium; and RV, right ventricle.

Increased productions of ANP (atrial natriuretic peptide) and BNP (brain natriuretic peptide) from atrial and ventricular tissue are markers of the increased chamber filling pressures in the setting of diastolic dysfunction.²⁵ Ranolazine treatment counteracted the increase in mRNA expression of both ANP- and BNP-coding genes (Table IV in the Data Supplement), in line with the prevention of diastolic abnormalities.

Cardiac Magnetic Resonance

A total of 8 WT, 8 R92Q-KET, and 8 R92Q-RAN mice underwent cardiac magnetic resonance tests performed using a 7-Tesla scanner (Bruker Pharmascan) optimized for small animals following a standardized protocol¹⁷ to measure chamber volumes and wall mass from sequential short-axis sections (Figure 4A). Consistent with the echocardiographic data, ranolazine prevented the reduction of LV end-diastolic volume and the increase of septal wall thickness and LV ejection fraction, which were observed in all untreated mutant mice (Figure 4B). Total LV wall mass was unchanged in R92QKET versus WT mice and was unaffected by ranolazine treatment (Figure 4B).

Figure 4.

Magnetic resonance. A, Representative short-axis magnetic resonance imaging (MRI) sections at mid-left ventricle (LV) at end diastole (left) and end systole (right) from wild-type (WT), R92Q-KET, and R92Q-RAN mice. Horizontal bars equal 1 mm. B, Measurements obtained from 3D reconstruction using 6 to 8 axial sections per heart. From top to bottom, LV volume at end diastole, LV ejection fraction (LV EF), diastolic septal thickness, and normalized LV mass (mg/mm of tibia length) in mice from the 3 study groups. Means±SEM from 9 mice per group. C, Representative gadolinium enhancement sections obtained 1 minute (left) and 30 minutes (right) after gadolinium injection in the tail vein in a WT mouse. D, Calculated fraction of extracellular space in WT, R92Q-KET, and R92Q-RAN hearts. Means±SEM from 8 mice per group. B and D, *P<0.05; **P<0.01; NS, P>0.05; statistical test: 1-way analysis of variance with Tukey correction.

A gadolinium compound was administered intravenously, and the decay of gadolinium contrast intensity in LV myocardium was used to estimate the fraction of extracellular space as an index of the interstitial fibrosis¹⁸ (Figure 4C). Ranolazine treatment significantly prevented the pathological expansion of extracellular volume in R92Q mutant mice (Figure 4D).

Histology and Confocal Imaging

Formalin-fixed LV samples were embedded in paraffin, and thin sections were stained with picrosirius red to highlight myocardial collagen fibers (Figure 5A). Analysis of stained sections from 5 animals from each of the 3 groups confirmed that lifelong ranolazine treatment counteracted the increase of myocardial collagen in R92Q mice (Figure 5B), in line with contrast magnetic resonance imaging data (Figure 4D). Myocardial fibrosis in cardiac diseases is subtended by increased local synthesis of transforming growth factor β.²⁶ Consistent with histological and magnetic resonance imaging data, ranolazine treatment prevented the increase in transforming growth factor β expression seen in R92Q mutants (Table IV in the Data Supplement).

Figure 5.

Myocardial structure studies. A, Representative histological sections of mouse left ventricular (LV) myocardium stained with hematoxylin (purple, marking cell nuclei) and picrosirius red (pink). The bright pink strands denote collagen fibers. Horizontal bar equals 100 μm. B, Fraction of the myocardium occupied by collagen fibers, expressed as percentage of the whole section’s area, measured in samples from mice of the 3 groups. Means±SEM from 9 mice per group. Four to 5 histological sections from each mouse were read to calculate the extent of picrosirius red positive areas. C, Representative confocal images from isolated LV cardiomyocytes stained with di-8-aneppdhq from wild-type (WT; top), R92Q-KET (middle), and R92Q-RAN (bottom) hearts. Horizontal bars equal 15 μm. D, T-tubule power, as calculated using the TTorg ImageJ plug-in, in cardiomyocytes from the 3 study groups. E, Mean T-tubule period (equivalent to sarcomere length) in resting LV cardiomyocytes from the 3 cohorts of mice. D and E, Means±SE from 72 WT (5 mice), 204 R92Q-KET (6 mice), and 269 R92Q-RAN (7 mice) cardiomyocytes. B, D, and E, **P<0.01; NS, P>0.05; statistical tests: Kruskal–Wallis test with Dunn’s multiple comparison test in B, Linear mixed models with Tukey–Kramer correction for multiple comparisons (LMMT) in D, LMMT adjusted for heteroscedasticity in panel E.

Freshly isolated mouse LV cardiomyocytes were stained with membrane-selective fluorescent dye di-3-aneppdhq (Life Technologies) and were then observed with a confocal microscope (Figure 5C) to assess the structure and organization of the T-tubular system.^22,27 The density of T-tubules and mean sarcomere length were then quantified using fast Fourier transform as previously described.²⁷ Ranolazine treatment partially prevented the reduction of the density and the impairment of the organization of T-tubules (TTPower parameter; Figure 5D) that were much more marked in cells from untreated R92Q mice. Of note, resting sarcomere length, which was significantly reduced in untreated R92Q cardiomyocytes, was the same in R92Q-RAN and WT cells (Figure 5E).

Intracellular Ca²⁺ Handling in Single Cardiomyocytes

All experiments in cells and trabeculae were performed using experimental solutions devoid of ranolazine. Intracellular Ca²⁺ measurements were performed in isolated cardiomyocytes during electric field stimulation at 3 different frequencies (1, 3, and 5 Hz). Figure 6 shows that ranolazine treatment prevented the alterations of Ca²⁺ handling observed in cardiomyocytes from mutant mice. Specifically, in R92Q-RAN cardiomyocytes, the increase in diastolic intracellular Ca²⁺ concentration ([Ca²⁺]_i) and its frequency-dependent rise observed in untreated mutant mice were markedly decreased (Figure 6A and 6B). Additionally, ranolazine treatment fully prevented the slowing of the kinetics of Ca²⁺-transient rise and decay (Figure 6A and 6C) and counteracted the reduction of Ca²⁺-transient amplitude (Figure 6A and 6D). In keeping with these observations, the reduction of Sarcoplasmic/Endoplasmic Reticulum Calcium ATPase type 2a (SERCA2a) protein and mRNA-expression of ATP2a2 gene, found in R92Q-KET hearts, was abolished in R92Q-RAN hearts (Table IV in the Data Supplement); this is likely to have contributed to the faster Ca²⁺ transient kinetics observed in R92Q-RAN versus R92Q-KET cardiomyocytes.

Figure 6.

Intracellular calcium in isolated cardiomyocytes and Western blot studies. A, Representative superimposed Ca²⁺ transients from R92-KET and R92Q-RAN myocytes elicited at 1 Hz (left), 3 Hz (middle), and 5 Hz (right). B, Diastolic Ca²⁺ levels, expressed as arbitrary units of fluorescence intensity, during steady state stimulation at different frequencies in cells from mice of the 3 cohorts. C, Time from stimulus to peak (peak time), time from peak to 50% decay (50% decay), and 90% decay of Ca²⁺ transients elicited at 1 Hz in cardiomyocytes from WT, R92Q-KET, and R92Q-RAN mouse hearts. D, Amplitude of Ca²⁺ transients in cardiomyocytes from the 3 study groups, elicited at stimulation rates of 1, 3, and 5 Hz. B–D, Means±SE from 125 WT (7 mice), 180 R92Q-KET (8 mice), and 378 R92Q-RAN (10 mice) cardiomyocytes. #P<0.01 for R92Q-KET vs WT and R92Q-RAN vs R92Q-KET at all frequencies; §P<0.01 at 1 Hz, P<0.05 at 3 Hz. E, Left, Superimposed representative fluorescence traces obtained during rapid exposure to 20 mmol/L caffeine after regular stimulation at 1 Hz, in WT, R92Q-KET, and R92Q-RAN cardiomyocytes. Right, Amplitude of caffeine-induced Ca²⁺ transients in mice from the 3 study groups. F, Kinetics of caffeine-induced Ca²⁺ transients (time from peak to 50% and 80% decay) measured from cardiomyocytes of the 3 study groups. E and F, Means±SE from 91 WT (4 mice), 90 R92Q-KET (4 mice), and 88 R92Q-RAN (4 mice) cardiomyocytes. G, Representative Western blots for total Ca²⁺–calmodulin-dependent protein kinase (CaMKII)δ, phospho-CaMKII at tyrosine 287, total SERCA2, total phospholamban (PLB), phospho-PLB at tyrosine 17 (CaMKII site), and Glyceraldehyde 3-phosphate dehydrogenase (GAPDH). H, Average values of WT, R92Q-KET, and R92Q-RAN hearts (8 samples each). Relative intensity of individual bands was quantitated and normalized to GAPDH. The ratio for control was assigned a value of 1. CAMK=CaMKII(δ) B–D, F and H, *P<0.05;**P<0.01; NS, P>0.05; statistical tests: Linear mixed models with Tukey-Kramer correction for heteroscedastic groups in B; LMMT in D–F; Welch’s analysis of variance (WA) in H.

Cardiomyocytes were rapidly exposed to 10 mmol/L caffeine after 30 s of conditioning stimulation at 1 Hz (Figure 6E) to estimate total sarcoplasmic reticulum Ca²⁺ content and study the rate of Ca²⁺ extrusion from the cytosol via the NCX. The amplitude of caffeine-induced Ca²⁺ transients was similar in WT and R92Q-KET cardiomyocytes and was not modified by treatment with ranolazine (Figure 6E), suggesting that no major changes in sarcoplasmic reticulum Ca²⁺ content occur in this disease model at 12 months of age. However, the drug prevented the slowing of the decay rate of caffeine-induced Ca²⁺ transients, suggesting that the rate of Ca²⁺ extrusion via the NCX is normalized in R92Q-RAN cardiomyocytes (Figure 6F). Interestingly, the mRNA expression of the NCX gene was similar in the 3 study groups (Table IV in the Data Supplement).

Western Blot Studies

Fast-frozen LV myocardial samples from R92Q, R92Q-RAN, and WT animals (8 per group) were processed to obtain total proteins, which were used for Western blot studies (Figure 6G and 6H), as previously described.¹⁰ CaMKII autophosphorylation, a marker of CaMKII activation, was increased in R92Q hearts as compared with that in WT mice. In ranolazine-treated R92Q mice, instead, the increase of CaMKII auto-phosphorylation was fully prevented. In agreement with this observation, ranolazine treatment prevented the increase of the phosphorylation of specific CaMKII target site on phospholamban. Additionally, ranolazine treatment prevented the reduction of total SERCA and phospholamban (PLB) proteins in R92Q mutant hearts. Phosphorylation of protein kinase A targets was similar in WT and R92Q hearts and remained unchanged by treatment (Figure IV in the Data Supplement). Finally, ryanodine receptor’s total protein amount and phosphorylation levels were found to be similar in the 3 groups (Figure IV in the Data Supplement).

Arrhythmogenic Ca²⁺ Release

We then assessed the rate of intracellular Ca²⁺ waves and spontaneous Ca²⁺ transients occurring during pauses. Toward this goal, cardiomyocytes were stimulated at 5 Hz, and pacing was abruptly stopped for 20 s, in the absence and presence of isoproterenol 10^–7 mol/L (Figure 7A). Treatment with ranolazine counteracted the increase of spontaneous Ca²⁺ release seen in untreated R92Q cardiomyocytes. Indeed, the frequencies of spontaneous Ca²⁺ waves and Ca²⁺ transients in cells from R92Q-RAN mice were markedly reduced compared with those from R92Q-KET mice, both at baseline and during β-adrenergic stimulation (Figure 7A and 7B), that is, they were in the range of those observed in cardiomyocytes from WT mice.

Figure 7.

Arrhythmogenic spontaneous calcium release. A, Representative traces showing the protocol to elicit Ca²⁺ waves in R92Q-KET (top) and R92Q-RAN (bottom) cardiomyocytes at baseline (left) and on administration of isoproterenol 10^–7 mol/L (Iso, right). Notably, R92Q-KET cardiomyocytes showed frequent Ca²⁺ waves (denoted by black arrows) and premature Ca²⁺ transients (blue arrows). Black lines below the traces indicate the times of stimulation. B, Occurrence of spontaneous Ca²⁺ waves (left) and spontaneous Ca²⁺ transients (right) in wild-type (WT), R92Q-KET, and R92QRAN cardiomyocytes at basal conditions and in the presence of isoproterenol (Iso). Means±SE from 106 WT (7 mice), 108 R92Q-KET (8 mice), and 240 R92QRAN (8 mice) cardiomyocytes. **P<0.01; statistical test: Linear mixed models with Tukey-Kramer correction.

Mechanical Function in Ventricular Trabeculae

Isometric force was measured from LV and right ventricular trabeculae or thin papillary muscles during field stimulation with different pacing protocols (Figure 8A and 8F). The response to stimulation frequencies (0.1–6 Hz) and the effects of β-adrenergic stimulation were evaluated, as well as the occurrence of spontaneous contractions during prolonged pauses (Figure 8B through 8G). Ranolazine treatment restored the positive inotropic responses that were markedly blunted in R92Q myocardium. In R92Q-KET trabeculae, we observed a significant reduction of twitch tension compared with that in WT, both at fast pacing rates (≥3 Hz) and in response to isoproterenol, while the positive inotropic effects in response to fast stimulation and isoproterenol in R92Q-RAN muscles were similar to those of WT (Figure 8B and 8C). Ranolazine treatment markedly mitigated the frequency-dependent increase of diastolic tension in R92Q trabeculae (Figure 8D). Moreover, the kinetics of contraction and relaxation at low pacing frequencies (0.1–1 Hz) were faster in R92Q-RAN trabeculae, as compared with those in vehicle-treated mutants (Figure 8E). Ranolazine, however, did not prevent the prolongation of twitch relaxation at higher frequencies (>1 Hz). Interestingly, treatment reduced the occurrence of spontaneous contractions during pauses to values similar to those observed in WT trabeculae (Figure 8F), a finding consistent with the reduction of intra-cellular Ca²⁺ waves and Ca²⁺ transients detected in isolated cardiomyocytes from R92Q-RAN hearts.

Figure 8.

Mechanics of intact and skinned trabeculae. A, Representative force traces from left ventricular (LV) trabeculae of wild-type (WT), R92Q-KET, and R92Q-RAN mice, stimulated at 4 Hz. B, Relationship between active twitch force and stimulation frequency (0.1–7 Hz) in trabeculae from the 3 groups. C, Twitch tension measured in the presence of isoproterenol 10^–7 mol/L (Iso). D, Diastolic (passive) tension measured during steady state stimulation at different frequencies (0.5–7 Hz). B–D, #P<0.05 for frequencies 1 to 7 Hz; §P<0.05 for frequencies 4 to 7 Hz. E, Time from stimulus to peak contraction (peak time, left) and time from peak to 50% relaxation (RT50%, right), measured during stimulation at different frequencies. ‡P<0.05 for frequencies 0.1 to 1 Hz; †P<0.05, 0.1 to 0.5 Hz; ║P<0.05, 0.1 to 3 Hz. F, Top,Representative traces from R92Q-KET (above) and R92Q-RAN (below) trabeculae displaying spontaneous beats during pauses. Gray lines indicate electric stimuli. Bottom, Percent of trabeculae showing spontaneous contractions during pauses (>2 in a 20 s pause) in intact trabeculae. Errors were calculated by approximating the binomial distribution with a normal distribution (central limit theorem). B–F, Means±SE from 9 WT (6 mice), 11 R92Q-KET (8 mice), and 9 R92Q-RAN (6 mice) trabeculae. G, Top, Average tension/pCa relationship in skinned trabeculae; bottom, pCa at 50% of maximal tension (pCa50). Means±SE from 5 WT (2 mice), 5 R92Q-KET (2 mice), and 5 R92Q-RAN (2 mice) trabeculae. C, F, G, *P<0.05; **P<0.01; NS, P>0.05; statistical tests: Linear mixed models with Tukey-Kramer correction in B–E; Welch’s analysis of variance in F and G.

Finally, as previously reported²³ the R92Q cTnT mutation is associated with a marked increase in myofilament Ca²⁺ sensitivity that was assessed in skinned ventricular trabeculae (Figure 8G). Ranolazine treatment did not affect the increase in myofilament Ca²⁺ sensitivity associated with the mutation. The pCa50 of force generation was higher in both untreated and treated R92Q skinned preparations versus WT and was similar in R92Q-KET and R92Q-RAN mice.

Discussion

R92Q-TnT Mouse Recapitulates the Main Features of Human HCM

In this work, we used a comprehensive, multilevel approach (from single cells to the whole heart) to demonstrate that the I_NaL blocker ranolazine slows and attenuates the development and progression of morphological and functional HCM phenotypes in mutant mice carrying the R92Q mutation in the cardiac TnT gene.

The mouse model we used in this study¹⁴ recapitulates the main pathophysiological features of HCM in patients, both at whole heart and single cardiomyocyte level. In particular, our echocardiographic measurements showed a marked diastolic dysfunction with LA enlargement in mutant mice (Figure 3), paralleled by slower twitch kinetics and increased diastolic tension of trabeculae (Figure 8). At single cell level, alterations in Ca²⁺ handling sub-tended such diastolic abnormalities, including delayed Ca²⁺ transient decay, diastolic Ca²⁺ accumulation, lower SERCA expression, and reduced NCX activity (Figure 6). Interestingly, increased I_NaL was a major contributor to the observed alterations of intracellular Ca²⁺ handling in R92Q cardiomyocytes. All these functional cellular abnormalities are qualitatively similar to those we previously observed in human HCM cardiomyocytes.¹⁰ The only remarkable difference between the R92Q mouse model and the human HCM cells is the absence of a prolonged ventricular AP in mouse myocytes, despite the observed increase of I_NaL. On the contrary, when the excess of I_NaL is inhibited by ranolazine, APD is significantly shorter in R92Q cardiomyocytes, as compared with WT cells. Previous work provided a possible explanation for this observation.^28,29 In a transgenic mouse model carrying the I79N mutation of cTnT (whose effects on sarcomeric function are similar to R92Q), the higher myofilament Ca sensitivity, by increasing the total cytosolic Ca-buffering capacity of the cardiomyocyte, led to lower Ca-transient amplitude. The reduced size of Ca²⁺ transients in I79N myocytes (as in R92Q cells, see Figure 6D and Supplementary Figure I in the Data Supplement), led to partial suppression of the Ca2⁺ extrusion mode of the exchanger, thereby, resulting in a smaller inward NCX-mediated current (I_NCX) during the repolarization phase, ultimately leading to shorter APD in mutant myocytes. An additional possible mechanism leading to AP shortening in R92Q cells is the activation of I_K,ATP channels because of reduction of ATP concentration caused by the reduced energy efficiency of mutant myocardium.¹⁵

In addition to the functional cellular abnormalities, the R92Q transgenic line showed a marked increase in endomyocardial fibrosis (Figures 4 and 5), a hallmark of HCM in patients, particularly in those carrying thin-filament gene mutations.⁶ Finally, a clear propensity to develop ventricular arrhythmias in response to adrenergic stimulation has been previously shown in this transgenic line³⁰ and was confirmed by our studies in single cells and intact trabeculae (Figures 7 and 8). The functional and structural remodeling of mutant myocardium was associated with increased activation of CaMKII, a nodal protein in the regulation of cardiac physiology and pathology, whose enhanced activity is notably associated with cardiac hypertrophy and heart failure.³¹ Interestingly, a similar increase in the phosphorylation of CaMKII and its downstream targets was also consistently found in cardiac samples from HCM patients.²¹ Of note, despite an increased septal thickness measured with echo and magnetic resonance imaging, total LV wall mass was similar in R92Q versus WT hearts, in line with earlier findings in mice from this lineage.¹⁴ HCM in patients is diagnosed in the presence of a maximal LV diastolic thickness >15 mm and not by the presence of an increased LV total mass.² Indeed, regional distribution of hypertrophy is common in HCM patients, and >35% of them show no or mild increase of total LV mass.³² Moreover, the absence of global LV hypertrophy and the restrictive physiology, far from representing a limitation of this model, are typical of thin-filament HCM and are often observed in patients carrying similar high-risk mutations.^5,6 All in all, the cardiac pathological phenotype observed in R92Q transgenic mice provides a better approximation of the wide spectrum of myocardial and cellular abnormalities observed in human pathology, as compared with other currently available HCM mouse models carrying mutations in Myosin Heavy Chain¹¹ or Myosin-binding Protein-C (MyBPC),³³ increasing the translational value of our study.

Rationale for the Use of Ranolazine as a Disease-Modifying Strategy

The choice of ranolazine as a lifelong preventive treatment was driven by the positive acute effects that this drug exerted in cells and trabeculae isolated from patients with HCM¹⁰ and by its safety profile and tolerability in patients. The mechanisms by which ranolazine is capable of normalizing intracellular Ca²⁺ are directly related to its effect to inhibit cardiac I_NaL. LV cardiomyocytes from R92Q hearts show a significantly increased density of I_NaL when compared with WT cells (Figure 1); interestingly, the increase of I_NaL is an early feature of cellular remodeling in R92Q myocardium, as it is already present in cells from 1-month-old mutant mice. Early hyperactivation of CaMKII (because of the onset of intracellular Ca²⁺ overload) increases the phosphorylation of NaV1.5 channels at specific sites determining impaired current inactivation and may, thus, be the cause of the early increase of I_NaL³⁴ in mutant hearts. Indeed, increased intracellular [Ca²⁺] is widely considered among the first common consequences of HCM-causing sarcomeric mutations. Increased myofilament Ca²⁺ sensitivity (Figure 8) leads to increased diastolic [Ca²⁺] as a consequence of the larger cytosolic Ca-buffering capacity.^28,29 The increased ATP consumption by the mutant sarcomere^15,35 may reduce the amount of ATP available to fuel SERCA-mediated Ca-reuptake into the sarcoplasmic reticulum, ultimately contributing to cytosolic Ca²⁺ overload, the main determinant of CaMKII activation and pathological I_NaL enhancement.

The increased I_NaL results into a marked intracellular Na⁺ accumulation, especially at high pacing rates (Figure 1), which in turn contributes to increase intracellular diastolic [Ca²⁺] by impairing the forward function of the NCX (Figure 6). As in human HCM,¹⁰ acute exposure to ranolazine, that markedly inhibited I_NaL, reduced both intracellular [Na⁺] and diastolic [Ca²⁺] in R92Q cardiomyocytes (Figure 1). The normalization of diastolic Ca²⁺ by ranolazine highlights the possibility of additional benefits in the long-term administration because increased diastolic Ca²⁺ accumulation is one of the main subcellular determinants of myocardial adverse remodeling and disease progressions in animal models of HCM irrespective of genotype.^11,36 Long-term treatment with ranolazine was effective in reducing the magnitude of cellular remodeling in other models of hypertrophy.^37,38 In spontaneously hyper-tensive rats, chronic ranolazine treatment started before the onset of hypertrophy prevented the development of T-tubules disruption and significantly attenuated the degree of alterations of Ca²⁺ handling.³⁷ In a model of deoxycorticosterone acetate salt–induced hypertension in rats, ranolazine treatment prevented the development of secondary modifications of the myocardial contractile apparatus, thus, reducing the severity of diastolic dysfunction.³⁸ However, long-term treatment with ranolazine has not been previously attempted in models of genetic heart disease.

Lifelong Ranolazine Treatment Prevents the Development of Diastolic and Systolic Dysfunction

In the present work, transgenic mice carrying the R92Q mutation of TnT were treated with ranolazine since birth and proved safe in this setting. Early initiation of treatment is likely essential because HCM-related alterations of myocardial structure and function are thought to develop early in life, possibly starting in the prenatal phase.^9,39 Ranolazine treatment effectively countered the development of diastolic dysfunction in our HCM mice, as evaluated by Doppler echocardiography (Figure 3); in keeping with the improved diastolic function, a lower diastolic tension was observed in trabeculae from treated mice (Figure 8). The effect of the treatment on diastolic function was associated with the prevention of pathological Excitation-Contraction-coupling changes: myocytes from treated R92Q mice lacked the increase in diastolic Ca²⁺ seen in vehicle-treated mice and showed a faster decay rate in Ca²⁺ transient because of maintained near-normal SERCA and NCX function. Preservation of SERCA function seems particularly relevant to the outcome of ranolazine treatment; indeed, SERCA overexpression via gene transfer was capable of reducing phenotypic expression and disease progression in a murine model of HCM carrying a tropomyosin mutation.³⁶

Although abnormalities of Ca²⁺ transient kinetics were fully prevented in R92Q-RAN cardiomyocytes, the prolongation of twitch relaxation kinetics was not entirely prevented in trabeculae from ranolazine-treated mice. The interplay between Ca²⁺ removal from the cytosol and the intrinsic ability of myofilaments to switch off contraction contributes to determine the velocity of twitch relaxation. In line with previous studies,¹⁴ skinned muscle from R92Q mice exhibited increased Ca²⁺ sensitivity (Figure 8). Because this feature is a direct consequence of the mutation altering the thin-filament structure⁴⁰ and occurs upstream from the site of action of ranolazine, it was not surprising that Ca²⁺ sensitivity was not affected by the treatment. Notably, this rules out any effect of ranolazine on transgene expression. The persistence of sarcomeric abnormalities explains why altered contraction kinetics are still present in R92Q-RAN myocardium, despite the normalization of intracellular Ca²⁺ dynamics.

The link between fibrosis and diastolic impairment is critically relevant for the pathophysiology of HCM, particularly for thin-filament mutations.⁶ Treatment with spironolactone (an aldosterone receptor blocker) and N-acetyl-cysteine (an antioxidant) attenuated fibrosis in transgenic mice carrying the R92Q-TnT mutation, and the reduction of fibrosis alone led to improvement in diastolic function.^41,42 It is, therefore, notable that ranolazine treatment counteracted the expansion of extracellular volume in the ventricular myocardium of R92Q mice, as evaluated with contrast magnetic resonance imaging, reduced the proliferation of intramyocardial fibrosis in histological sections (Figures 4 and 5), and prevented the increase of transforming growth factor β expression.

In addition to preventing diastolic dysfunction, lifelong ranolazine treatment also prevented the development of systolic abnormalities and rescued the blunted inotropic responses in mutant mice. Ventricular trabeculae from R92Q mice showed a slight reduction in twitch amplitude under baseline conditions but displayed significantly reduced responses to a variety of positive inotropic challenges mimicking exercise conditions, including high-frequency stimulation and β-adrenergic activation with isoproterenol. The reduction of inotropic reserve in R92Q mice was completely abolished by ranolazine (Figure 8). The clinical relevance of this effect is also important because progression toward systolic dysfunction and the so-called end stage occurs in 8% to 10% of patients with HCM and in ≤20% of patients carrying thin-filament mutations.⁶ Of note, ranolazine prevented the disruption and disorganization of T-tubules, as evaluated with confocal microscopy using fluorescent membrane staining (Figure 5), an effect that likely contributed to the preservation of LV diastolic and systolic function.²²

Ranolazine Treatment Prevents the Development of Arrhythmogenic Substrates

Our results suggest that long-term ranolazine treatment is able to counteract the development of an arrhythmogenic substrate in the myocardium of mutant mice with HCM. Although we did not directly evaluate the frequency of ventricular ectopy nor the inducibility of arrhythmias in living animals, cardiomyocytes from treated mice showed a marked reduction in frequency of Ca²⁺ waves and spontaneous Ca²⁺ transients during pauses, both at baseline and during challenge with isoproterenol. Likewise, spontaneous contractions in intact trabeculae from R92Q-RAN mice were nearly absent, while they occurred with relatively high frequency in vehicle-treated mice. The reduction of Ca²⁺-dependent arrhythmogenesis is likely subtended by the normalization of intracellular Ca²⁺ handling by ranolazine,¹⁰ although other factors, such as the prevention of transcriptional changes in ion channel genes, such as I_to, may have contributed to the effect of the treatment (Table IV in the Data Supplement). CaMKII can acutely regulate ion channels (I_Na, I_Ca, I_K) and Ca²⁺-handling proteins (RyR, IP3R, PLB), directly contributing to triggered arrhythmias (such as Early After-depolarizations and Delayed After-depolarizations).⁴³ Reduction of CaMKII activity by ranolazine treatment is likely to have contributed to the reduction of arrhythmias in treated R92Q mice. Finally, at the whole heart level, the reduction of cellular arrhythmic triggers is paralleled by a reduced intramyocardial fibrosis (a substrate for re-entry circuits), potentially lowering the risk of sustained ventricular arrhythmias.

Mechanisms Underlying the Efficacy of Ranolazine in R92Q-TnT Mice

The CaMKII-I_NaL vicious cycle is a well-established phenomenon occurring in several cardiac diseases, including human HCM. In brief, increased I_NaL leads to intracellular Na⁺ overload, which impairs NCX-mediated extrusion of Ca²⁺, thus, contributing to Ca²⁺ overload, which in turn potentiates CaMKII activity via calmodulin binding¹³; in addition, Na⁺ overload increases mitochondrial reactive oxygen species production and oxidation of CaMKII, leading to its constitutive activation.⁴⁴ The hyperfunctional CaMKII, via direct phosphorylation of NaV1.5 Na⁺ channels, contributes to further increase in I_NaL.³⁴ Although increased I_NaL is never the primary cause of disease (with the exception of LQT3), it is a common feature of cardiac cell remodeling in several conditions, including mice carrying the R92Q TnT mutation (Figure 1). Inhibition of I_NaL is likely to thwart the establishment of the aforementioned vicious cycle, leading to a sustained reduction of CaMKII activity. Indeed, we have found that the degree of CaMKII autophosphorylation at Thr286 site is markedly increased in untreated R92Q mice, while it is undistinguishable from WT in ranolazine-treated mutants (Figure 6G and 6H). Prolonged CaMKII hyperactivity in cardiac diseases is associated with changes of gene expression program that ultimately contribute to drive myocardial remodeling both at cellular level (hypertrophy, changes of cellular substructures, expression of ion channels, and EC-coupling proteins) and at extracellular level (intramyocardial fibrosis). Activated CaMKII can migrate to the nucleus and phosphorylate hystone deacetylase (HDAC4), thus, relieving the inhibition of the transcription of myocyte enhancer factor-2 (MEF2)-controlled genes.⁴⁵ In addition, Ca-activated calmodulin may also bind to calcineurin and dephosphorylate Nuclear factor of activated T cells (NFAT), which in turn may increase GATA-dependent transcription.⁴³ Interestingly, MEF2 and GATA-controlled genes are involved in the development of pathological cardiomyocyte hypertrophy, as well as in the often associated extracellular matrix remodeling and fibrosis. Therefore, the reduction of CaMKII activity after I_NaL inhibition in ranolazine-treated R92Q mice is likely to be the main mechanism that prevented disease-related structural and functional myocardial changes in treated mice.

In a recent work by Flenner et al,⁴⁶ mice carrying an MyBPC mutation related to HCM were treated with ranolazine, and most of the observed effects of the drug in that model were found to be related with a significant inhibition of β-adrenergic receptors. Ranolazine has a significant β-blocking effect when the plasma concentration is >13 μmol/L,⁴⁷ much higher than the clinically relevant concentration range (from 2 to 5 μmol/L). Indeed, Flenner et al⁴⁶ reached very high plasma ranolazine concentrations in their treated mice (>20 μmol/L), so a β-blocking effect observed with their protocol is expected, but scarcely clinically relevant. In our ranolazine-treated male mice, plasma ranolazine concentration ranged from 2.5 to 5.5 μmol/L (average 4 μmol/L). Moreover, they do not show lower heart rate and blood pressure when compared with vehicle-treated mice (Table II in the Data Supplement). Furthermore, phosphorylation of protein kinase A targets is not reduced in the hearts from treated versus untreated R92Q mice (Figure IV in the Data Supplement). Therefore, we can reliably exclude that a significant β-blocking action plays a role in determining the effects of ranolazine in R92QTnT mice. Finally, despite their widespread use in HCM patients, β-blockers have never been proven to be capable of slowing or preventing the development or progression of disease in murine HCM models or in patients^3,7 on long-term administration.

Implications for Clinical Management

Taken together, our data support the role of ranolazine or other I_NaL inhibitors to prevent, delay, or attenuate the onset and severity of disease in subjects carrying mutations associated with severe forms of HCM.^5,6 The feasibility of preventive treatment strategies in young HCM mutation carriers has been recently ascertained by Ho et al,¹² who recently completed a trial with diltiazem and are currently performing a trial with valsartan in this group of patients (VANISH Trial [Valsartan for Attenuating Disease Evolution In Early Sarcomeric HCM]).

Although the present results support the use of ranolazine in subjects carrying high-risk thin-filament mutations, whether its efficacy extends to other genotypes remains to be determined. A recent study by Flenner et al⁴⁶ showed that a 6-month treatment with ranolazine did not reduce cardiac hypertrophy or diastolic dysfunction in an MyBPC-mutant HCM mouse model. The lack of effects of ranolazine in that work stems from the absence of a significant increase of I_NaL in cardiomyocytes from MyBPC-mutant mice as compared with WT littermates,⁴⁶ at variance with the R92Q-TnT model. In human cardiomyocytes from adult patients with HCM, however, we previously showed that similar electric abnormalities (including increased I_NaL) are also present in cardiomyocytes from patients carrying MyBPC mutations,¹⁰ and the benefits of acute ranolazine administration seemed independent from the presence of a specific mutation. The cardiac pathological phenotype of the MyBPC mouse model in the work by Flenner et al⁴⁶ was rather mild, with slight diastolic dysfunction, minimal LV hypertrophy, and little cellular abnormalities,³³ and provided a lesser approximation of advanced human disease as observed in patient samples,¹⁰ as compared with the R92Q model. Moreover, Flenner et al started treatment in young adult mice of 2 months of age when disease phenotype was already fully expressed,³³ while we initiated treatment right after birth, before HCM phenotypic expression. This suggests that initiation of treatment before the onset of cardiac pathological changes is preferable to achieve a disease-modifying effect in HCM. Indeed, other supposedly disease-modifying agents, such as sartans or statins, have failed to reduce established hypertrophy and fibrosis in adult HCM patients with overt disease,⁴⁸ while early treatment of young mutation carriers with diltiazem has somewhat succeeded.¹²

Based on the results of the present work, we think that it would be essential for a preventive treatment with I_NaL inhibitors to be started before the onset of pathological phenotype, that is, during early to late childhood (5–14 years). We do not think that initiation of treatment at birth in humans would be necessary to achieve the expected results. The reason why we decided to treat cTnT-R92Q mice since birth is because pathological cardiac changes (including increased I_NaL) start developing early in R92Q mice and are mostly preset at 1 month of age. Because it was impossible for us to define the exact time of the onset of disease-related changes during the first month of life, we decided to initiate treatment right after birth. Comparing the development of HCM in transgenic mice with that occurring in human children and young adults is nearly impossible, given the remarkable differences in the physiology of heart maturation and growth. At birth, the functional maturity of cardiomyocytes in the mouse heart is relatively higher than that in newborn humans.⁴⁹ From a translational standpoint, we think that starting treatment in a newborn mouse would be the equivalent of starting treatment in a 6-year-old child. This strategy could be feasible in individuals carrying high-risk mutations, such as thin-filament mutations, identified in familial screening programs.

Our work highlights the role of ranolazine as an agent to treat young carriers of high-risk mutations (eg, thin-filament mutations or aggressive MYH7 mutations). These subjects have a relatively high risk of lethal arrhythmias during adolescence and early adulthood and an increased likelihood of developing terminal heart failure later in life, with either hypocontractile or restrictive phenotype.^5,6 These patients are currently orphan of effective therapies because current strategies are unable to prevent arrhythmias or heart failure; thus, early prevention of phenotype development is the only possible therapeutic option.

Ranolazine is already available in the clinic for angina treatment and has been recently assessed in patients with symptomatic HCM (RESTYLE-HCM study [Ranolazine in Symptomatic Hypertrophic Cardiomyopathy]; EudraCT number, 2011-004507-20); in addition, eleclazine, a novel selective I_NaL inhibitor with improved efficacy and safety, is currently under clinical development and is being trialed in HCM patients (LIBERTY-HCM study [Effect of Eleclazine (GS-6615) on Exercise Capacity in Subjects With Symptomatic Hypertrophic Cardiomyopathy]; EudraCT number, 2013-004429-97).⁵⁰ The results of these studies may provide clinical evidence in support to our observations and answer the question whether the strategy of I_NaL inhibition is effective in ameliorating the clinical profile of adult patients with overt HCM. Whatever the results of these studies, however, the present work strongly supports a rationale for ranolazine or other I_NaL inhibitors as promising preventive therapy in carriers of HCM mutations at risk of developing a severe form of the disease.

CLINICAL PERSPECTIVE.

Family screening programs are detecting increasing numbers of young individuals who carry mutations associated with hypertrophic cardiomyopathy (HCM) before the onset of cardiac hypertrophy and dysfunction. Recent advances in HCM patient phenotyping led to the identification of several high-risk mutations (eg, thin-filament mutations) associated with early sudden death and progression toward heart failure. To date, however, when a young individual is diagnosed with having a high-risk mutation, clinicians can only suggest a close follow-up strategy, for lack of robust therapeutic options capable of preventing disease onset and progression. Although gene therapy still seems distant, an attractive alternative is represented by disease-modifying drugs interfering with the cellular mechanisms of myocardial adverse remodeling. Our previous studies in surgical samples from HCM patients showed that intracellular calcium overload and increased late-sodium current are key determinants of the electric and mechanical dysfunction of HCM myocardium, which were reverted by the late-sodium current inhibitor ranolazine. In a clinically relevant mouse model carrying the high-risk R92Q mutation of TnT gene, we here show that lifelong treatment with ranolazine, initiated before the onset of cardiac phenotype, is capable of preventing structural and functional cardiac changes associated with HCM (LV hypertrophy, intramyocardial fibrosis, diastolic dysfunction, and arrhythmogenesis). In ranolazine-treated mice, reduction of calcium overload prevents the activation of calmodulin kinase 2, thereby stopping the pathway of myocardial hypertrophic remodeling. Ranolazine is currently in clinical use for angina and has a remarkable safety profile. The present work supports a rationale for the rigorous evaluation of ranolazine as preventive therapy in young carriers of high-risk HCM mutations.