Targeting the sarcomere in inherited cardiomyopathies

Abstract

Variants in >12 genes encoding sarcomeric proteins can cause various cardiomyopathies. The two most common are hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM). Current therapeutics do not target the root causes of these diseases, but attempt to prevent disease progression and/or to manage symptoms. Accordingly, novel approaches are being developed to treat the cardiac muscle dysfunction directly. Challenges to developing therapeutics for these diseases include the diverse mechanisms of pathogenesis, some of which are still being debated and defined. Four small molecules that modulate the myosin motor protein in the cardiac sarcomere have shown great promise in the settings of HCM and DCM, regardless of the underlying genetic pathogenesis, and similar approaches are being developed to target other components of the sarcomere. In the setting of HCM, mavacamten and aficamten bind to the myosin motor and decrease the ATPase activity of myosin. In the setting of DCM, omecamtiv mecarbil and danicamtiv increase myosin activity in cardiac muscle (but omecamtiv mecarbil decreases myosin activity in vitro). In this Review, we discuss the therapeutic strategies to alter sarcomere contractile activity and summarize the data indicating that targeting one protein in the sarcomere can be effective in treating patients with genetic variants in other sarcomeric proteins, as well as in patients with non-sarcomere-based disease.

Graphical Abstract

Variants in genes encoding sarcomeric proteins can cause hypertrophic or dilated cardiomyopathy. In this Review, the authors discuss therapeutic strategies to target the cardiac sarcomere, focusing on four small molecules that have been developed that inhibit or activate the myosin motor protein to decrease or increase contractile force, respectively.

Introduction

In 1990, the first genetic variant linked to hypertrophic cardiomyopathy (HCM) was discovered in MYH7, which encodes cardiac myosin heavy chain-β¹. Since then, HCM has been called a ‘disease of the sarcomere’ because hundreds of variants in genes encoding cardiac myosin and other sarcomeric proteins have been linked to the disease. Similarly, variants in genes encoding sarcomeric proteins were implicated in genetic dilated cardiomyopathy (DCM) in 2000². In a major advance, genetic variants in TTN, encoding titin, were identified in 30% of patients with genetic DCM³. Compared with HCM, DCM can be caused by variants in a wider range of genes. In addition to variants in genes encoding sarcomeric proteins, variants in genes encoding non-sarcomeric proteins (including nuclear envelope proteins, transcription cofactors and cytoskeletal proteins) can cause DCM⁴. The mechanisms of how variants in genes encoding sarcomeric proteins cause disease have been studied in detail⁵, and some of these findings have been informative in defining and refining how small molecules that target the sarcomere are likely to change the course of clinical disease.

Clinically, HCM is characterized by a reduced left ventricular cavity (and thickened left ventricular walls), with hyperdynamic contraction, diastolic dysfunction and increased energy consumption^6,7. Surgical interventions in patients with HCM include implantation of a cardioverter–defibrillator, heart transplantation and, in those with obstructive HCM, surgical myectomy⁸. In some patients with obstructive HCM, alcohol septal ablation can be used instead of surgical myectomy⁹. Medical management for patients with HCM includes the use of β-blockers, Ca²⁺-channel blockers and, in those with symptomatic, obstructive HCM, the antiarrhythmic drug disopyramide because of its negative inotropic effects¹⁰. Despite medical and surgical management, HCM is a progressive disease, with atrial fibrillation and heart failure occurring in 20% of patients by the age of 50–70 years⁶.

By contrast, DCM is characterized by thinned ventricular walls and reduced contractility⁷. Treatments for patients with DCM include symptom management and approaches to preserve myocardial function. Effective medical therapies include mineralocorticoid-receptor antagonists, neprilysin inhibitors and sodium–glucose cotransporter 2 inhibitors, and cardiac resynchronization therapy can be used in a subset of patients^7,11. When medical management is insufficient to prevent disease progression, patients can develop heart failure and require heart transplantation or implantation of a left ventricular assist device. Therefore, an unmet need exists to modulate cardiac contractility directly without systemic effects on other organs; particularly in the setting of cardiomyopathy caused by variants in genes encoding sarcomeric proteins. Although the distinct clinical phenotypes of HCM and DCM require different therapeutic strategies, drug development for both has converged on the modulation of cardiac myosin function¹².

Manipulating the myosin motor directly (either increasing or reducing its function) has been the subject of intense interest over the past decade, culminating in several successful clinical trials for patients with HCM or DCM. The strategy to identify myosin modulators was to screen for molecules that bind to cardiac myosin specifically, alter its ATPase activity in vitro, and increase or decrease contractility in vivo. To ensure cardiac selectivity, candidate compounds were screened against non-muscle, smooth muscle or fast-twitch skeletal myosins. Of note, although many sarcomeric protein isoforms in the heart are different from those in skeletal muscle, the MYH7 gene product (known as the myosin heavy chain-β or slow-twitch isoform) is also expressed in type I fibres in skeletal muscle. Therefore, small molecules that modulate myosin heavy chain-β in the heart should also modulate myosin in slow-twitch skeletal muscle, given that it is the same protein^13,14.

In this Review, we discuss therapeutic strategies to target the cardiac sarcomere. Progress has also been made in therapeutic approaches to modulate the skeletal muscle sarcomere¹², but that is beyond the scope of this article. We discuss how targeting one protein in the sarcomere can be effective in treating patients with genetic variants in other sarcomeric proteins, as well as in patients with genetic variants in non-sarcomeric proteins.

The cardiac sarcomere: an overview

The two main filaments in sarcomeres are the Ca²⁺-dependent, regulatory thin filaments and the myosin-based, force-generating thick filaments¹⁵ (Fig. 1). The thin filament is composed primarily of five proteins: three that form the troponin complex (troponin C, the Ca²⁺-binding subunit; troponin I, the inhibitory subunit; and troponin T, the tropomyosin-binding subunit), tropomyosin and actin. These proteins work together in a highly coordinated fashion to transmit the Ca²⁺ status of the myofilament, which, in turn, regulates interactions between the thin and thick filaments¹⁶. Parallel to the thin filament, the thick filament is comprised of myosin molecules, which couple the chemical energy of ATP hydrolysis to mechanical movement (shortening) of the sarcomere. Importantly, myosin ATPase activity is highly coordinated with Ca²⁺ activation of the thin filament to ensure actomyosin interactions and force generation.

Fig. 1 ∣. The cardiac sarcomere and its components.

The cardiac sarcomere is made up primarily of regulatory thin filaments, force-generating thick filaments (myosin), regulatory myosin-binding protein C and tension-sensing titin. Variants in the genes encoding these proteins are linked to various genetic cardiomyopathies, such as hypertrophic cardiomyopathy (HCM) and dilated cardiomyopathy (DCM), which are typically characterized by a unique molecular mechanism of disease. Therefore, these proteins are potential druggable targets for treating cardiomyopathies.

The ATPase activity of myosin is defined by three distinct rates: actin-activated activity (fast), the disordered relaxed (DRX) state (~100-fold slower than actin-activated) and the super-relaxed (SRX) state (~10-fold slower than DRX) (Fig. 2). Myosin heads that are not bound to actin exist in a dynamic equilibrium between the DRX and SRX states^17-19. The DRX state is defined by the basal ATPase activity of myosin, in which the myosin heads are released from the thick filament backbone and available for interaction with actin. The SRX state is defined by an ultra-slow, energy-conserving ATPase activity of myosin^20-22. This reduction in ATPase activity is, in part, achieved by a structural conformation of the myosin molecule, termed the interacting heads motif (IHM), in which the motor domain is folded back onto the thick filament axis, sequestering the myosin in an auto-inhibited state^23-26. From electron microscopy images, 3D models of the IHM were mapped to mouse ventricular thick filaments, suggesting the existence of the IHM structural state in cardiac tissue, although the images were low resolution¹⁹. Another study measured the IHM in bovine cardiac heavy meromyosin via FRET (fluorescence resonance energy transfer), providing higher-resolution structural confirmation that the IHM structure does exist in cardiac muscle²⁶. However, the researchers note that perturbations to the SRX state did not produce concomitant alterations to the IHM population, suggesting that the SRX state in cardiac muscle is not always synonymous with IHM; instead, additional energy-conserving myosin conformation(s) are likely to exist^27,28.

Fig. 2 ∣. The functional states of myosin.

a ∣ Myosin molecules exist in an equilibrium between three distinct ATP-utilization states: the actin-bound state (green), the disordered relaxed (DRX) state (orange) and the super-relaxed (SRX) state (red). The SRX state can be divided into the folded back, interacting heads motif (IHM) structure or by another, undefined structural state. These states are defined by specific myosin ATPase activities, in which the activity in the actin-bound state is ~100-fold faster than in the DRX state and nearly 1,000-fold faster than in the SRX state⁸³. These differences in ATPase activity govern ATP utilization of the sarcomere and the heart. b ∣ The simulated ATPase activity for each state is shown as a function of time. The black curve depicts the myosin ATPase activity of cardiac muscle based on the physiological distribution of myosin heads between the three states. Adapted with permission from REF.¹³⁹.

In addition to the thin and thick filaments, cardiac sarcomeres contain titin and myosin-binding protein C (MyBP-C), two large proteins that regulate the function of the sarcomere (Fig. 1). Titin, the largest protein in the human proteome, spans the half sarcomere and is crucial in maintaining sarcomere structure throughout muscle contraction and relaxation²⁹. MyBP-C has been shown to interact with both the thin and thick filaments and titin. MyBP-C has been postulated to have a primary role in the regulation of myosin–actin cross-bridge formation^28,30-33, but the exact mechanisms of how MyBP-C regulates sarcomere function are incompletely understood. Interestingly, MyBP-C has been shown to regulate the SRX state in both in vitro systems and animal models^28,34-37. In skeletal muscle, the population of SRX myosin heads was higher in the so-called C-zone (where MyBP-C resides) than in the other thick filament zones³⁵. Loss of MyBP-C led to destabilization of the SRX population in mouse isolated left ventricular cardiomyocytes³⁴, and protein kinase A phosphorylation of MyBP-C destabilized the SRX population in mouse hearts³⁸, suggesting that transient recruitment of myosin heads out of the SRX state is, in part, regulated by MyBP-C.

Therapies for different aetiologies

When targeting the sarcomere to treat inherited cardiomyopathies, it is important to consider the different mechanisms of pathogenesis for each target. Given that the vast majority of genetic cardiomyopathies are inherited in an autosomal dominant fashion, therapeutic strategies for recessively inherited cardiomyopathies are not discussed in this Review. In addition, different strategies might need to be developed if the disease is the result of haploinsufficiency caused by either loss of function (via a reduction in the amount of functional protein) or gain of function (via a dominant-negative polypeptide mechanism). Finally, genome-wide association studies have identified a growing number of cardiomyopathy-associated variants in genes encoding non-sarcomeric proteins³⁹. Patients with these genetic variants were included in the clinical trials discussed below; however, the mechanism of action for sarcomere-targeted therapies in patients with variants in non-sarcomeric proteins is not yet well understood.

Most variants in MYH7 (the second most common cause of HCM) are autosomal dominant and largely result in single amino acid changes⁴⁰. Cardiomyopathy is caused by the mutant protein interfering with the function of the wild-type protein. HCM-causing variants in genes encoding most other sarcomeric proteins, such as the troponin subunits and tropomyosin isoforms, also fall into this gain-of-function category, which is unsurprising in a structure such as the sarcomere, in which many different proteins assemble into structures that function together.

By contrast, the majority of variants in MYBPC3 (the most common cause of genetic HCM) cause frameshift mutations that result in a reduction in the total quantity of functional protein⁴¹. The reduced levels of functional protein could be due to a reduction in the amount of mRNA owing to nonsense-mediated decay of the mutant mRNA or, theoretically, to the production of truncated, inactive protein from the mRNA carrying the frameshift mutation. This mechanism of pathogenesis would be haploinsufficiency caused by a loss-of-function mechanism. For a long time, MYBPC3 mRNA that carried a frameshift mutation was proposed to encode a truncated protein that functioned as a poison polypeptide, which would be a gain-of-function mechanism⁴². However, no convincing data indicate that truncated forms of MyBP-C are translated from the mRNA carrying the frameshift mutation, and haploinsufficiency caused by loss of function is now widely considered to be the mechanism of most disease caused by MYBPC3 variants. Analyses of cardiac samples from patients with MYBPC3 variants suggest that the wild-type allele can compensate to some extent; the amount of wild-type MyBP-C protein was ~70% of normal levels, rather than the predicted 50%⁴¹. This finding is still consistent with a mechanism of haploinsufficiency if, for example, a threshold of >70% MyBP-C expression needs to be exceeded to have a normally functioning heart. Using induced pluripotent stem cell-derived cardiomyocytes from individuals who were heterozygous for MyBP-C-truncating variants, the same researchers found that the compensation by the wild-type allele was sufficient to achieve no decrease in total MyBP-C protein levels and no contractile deficits⁴³. Furthermore, the compensation seemed to be the result of decreased degradation of the wild-type MyBP-C protein⁴³. However, these observations in cardiomyocytes derived from induced pluripotent stem cells in vitro might not be relevant to an intact heart in vivo.

Importantly, the hypothesis that destabilization of the SRX state of myosin is a primary molecular driver of the hyperdynamic contractile phenotype seen in patients with HCM has been investigated⁴⁴. Biophysical and biochemical studies have demonstrated that single-nucleotide variants in MYH7 can increase the DRX state of myosin, leading to increased contractility and energy expenditure in the heart^22,37,45-50. Therefore, regulation of the SRX state is a potential mechanistic avenue for myosin-targeted small molecules (discussed below).

Variants in the same gene (such as MYH7 or TNNI3) can lead to either HCM or DCM⁵¹. The mechanisms whereby two different variants in the same gene lead to two different diseases are not well understood, although the most logical hypothesis is that different variants result in different amino acid substitutions, which affect the protein function differently. However, disease phenotype could also be affected by genetic background. For example, variants in LTBP4 can modify phenotypes in patients with Duchenne muscular dystrophy⁵². Furthermore, a missense variant in TTN has been shown to modify the phenotype of patients with a variant in LMNA⁵³.

The discovery that TTN variants are common causes of genetic DCM has led to important discoveries not only about the pathophysiology of DCM but also about titin biology^3,54. Approximately 30% of DCM can be attributed to variants in TTN³. The vast majority of these variants are inherited in an autosomal dominant manner, and many of them are called truncating variants because they have the potential to produce a truncated titin protein if transcribed and translated⁵⁵. However, in contrast to disease-causing variants in MYBPC3, multiple mechanisms of pathogenesis are associated with TTN variants, including both haploinsufficiency and truncated titin polypeptides as well as variants that alter post-translational modifications of titin⁵⁵. In 2021, definitive proof of the presence of truncated titin polypeptides in patients with DCM was reported⁵⁶.

Sarcomeric contractile activators

Small molecules targeting the cardiac sarcomere to increase contractility have largely been developed to treat idiopathic heart failure but could, in principle, ameliorate inherited cardiomyopathies characterized by poorly contracting hearts. These activators directly increase the Ca²⁺ sensitivity of the thin filament or increase the ATPase activity of the myosin motor. In this section, we discuss the preclinical and clinical data that are available for compounds that increase sarcomeric contractility and cardiac function.

Ca²⁺ sensitizers: targeting thin filaments

Therapy to increase cardiac contractility in hypocontractile diseases (such as heart failure) has for a long time substantially relied on three types of inotropic agent: digitalis-derived cardiotropic glycosides, the oldest inotropic drugs, which inhibit the sodium–potassium ATPase pump, resulting in the accumulation of Na⁺ that, in turn, reverses the sodium–calcium exchanger and leads to increased Ca²⁺ influx in systole; β-adrenergic agonists, such as dopamine or dobutamine; and phosphodiesterase inhibitors, which lead to increased levels of cAMP, thereby avoiding desensitization and downregulation of cardiac β-adrenergic receptors. All these agents rely on a generalized disruption of Ca²⁺ homeostasis in cardiomyocytes to increase contractility⁵⁷. However, the long-term use of β-adrenergic agonists or phosphodiesterase inhibitors has adverse effects, including maladaptive cardiac remodelling, probably resulting from the chronic activation of protein kinase A and Ca²⁺/calmodulin-dependent protein kinase II signalling. In the 1980s, a new strategy was proposed, aimed at directly increasing the Ca²⁺ sensitivity of the sarcomere by targeting the myofilaments^58-62. Independent of the mechanism of action, Ca²⁺ sensitizers induce a leftward shift in the curve describing the relationship between contractile force and Ca²⁺ concentration, leading to increased systolic force generation for any given cytoplasmic concentration of Ca²⁺. Conversely, Ca²⁺ sensitizers inherently impair diastole by slowing relaxation when the cytoplasmic Ca²⁺ concentration decreases. Levosimendan, a Ca²⁺ sensitizer that does not prolong relaxation^61,62, stabilizes the Ca²⁺-bound troponin C conformation and increases Ca²⁺ sensitivity^63-65. Additionally, levosimendan increases intracellular concentrations of cAMP⁶⁶ via a potent and selective inhibition of phosphodiesterase type 3^60,67,68 Therefore, levosimendan is likely to act via a combination of increasing Ca²⁺ sensitivity by binding to cardiac troponin C and ‘classic’ phosphodiesterase inhibition that increases Ca²⁺ sensitivity without impairing relaxation. Numerous clinical trials have been conducted to investigate the effects of levosimendan in patients with heart failure. Initial findings from the LIDO trial⁶⁹ suggested short-term beneficial effects of levosimendan in patients with acutely decompensated chronic heart failure. However, levosimendan had no significant effect on all-cause mortality or other secondary outcomes (cardiovascular mortality or number of days alive and out of hospital) in patients with chronic heart failure in the SURVIVE trial^69,70 or in patients with heart failure undergoing cardiac surgery in the CHEETAH⁷¹, LEVO-CTS⁷² and LICORN⁷³ trials (Table 1). Furthermore, in the REVIVE trial⁷⁴, levosimendan treatment was associated with an increased risk of adverse cardiac events, including increased heart rate and incidence of cardiac arrhythmia and frequent occurrence of hypotension in patients with acutely decompensated chronic heart failure. The ESC heart failure guidelines recommend levosimendan treatment in the acute management of patients with hypotension and/or symptoms of hypoperfusion⁷⁵; however, levosimendan is not approved for use in the treatment of heart failure in the USA.

Table 1 ∣.

Clinical trial results of sarcomere-targeted small molecules

Small molecule	Sarcomeric target	Molecular mechanism	Contractile effect	Clinical trial	Clinical effects	Refs
Sarcomeric contractile activators
Levosimendan	Troponin C (and PDE3)	Stabilizes troponin I–troponin C interaction	Increased owing to increased Ca2+ sensitivity	SURVIVE, REVIVE, CHEETAH, LEVO-CTS	Improved haemodynamic profile, no effect on cardiac relaxation; increased incidence of cardiac arrhythmia, no improvement in heart failure outcomes or mortality	59-74
CK-136 (AMG-594)	Troponin complex	Increases Ca2+ sensitivity	Increased owing to increased Ca2+ sensitivity	–	Not available	76,77
Omecamtiv mecarbil	Myosin heavy chain-β	Stabilizes the pre- powerstroke state of myosin	Increased owing to increased actomyosin interactions	GALACTIC-HF	Modest reduction in the rate of heart failure events and cardiac death; greatest improvement in function in patients with a left ventricular ejection fraction ≤28%	92,94-95
				COSMIC-HF	Improved left ventricular systolic ejection time, stroke volume, end-diastolic diameter and volume, and heart rate; reduced plasma NT-proBNP level	89, 90
Danicamtiv (MYK-491)	Myosin heavy chain-β	Unpublished	Increased	Phase II trial ongoing, recruiting participants with dilated cardiomyopat hy	Results not available	99,100
Sarcomeric contractile inhibitors
Mavacamten (MYK-461)	Myosin heavy chain-β	Stabilizes the super-relaxed (SRX) state of myosin	Decreased owing to reduced myosin head availability	EXPLORER-HCM	Improved peak oxygen consumption and NYHA functional class; reduced left ventricular outflow tract obstruction	118-120
				PIONEER-HCM	Improved post-exercise peak left ventricular outflow tract gradient	117
				MAVERICK-HCM	Decreased risk of adverse events, decreased plasma NT pro-BNP and cardiac troponin I levels	116,122
Aficamten (CK-274)	Myosin heavy chain-β	Slows phosphate release from myosin	Decreased owing to stabilization of weak actin-binding myosin conformation	REDWOOD-HCM	Results not available	135

NT-proBNP, N-terminal pro-B-type natriuretic peptide; PDE3, phosphodiesterase type 3.

In March 2020, Amgen and Cytokinetics presented preclinical data on CK-136 (formerly referred to as AMG-594) showing that CK-136 is a selective activator of the cardiac troponin complex, increasing sarcomere Ca²⁺ sensitivity and resulting in an increased number of myosin–actin cross-bridges and therefore higher contractile force⁷⁶. Preliminary data from studies using rat isolated ventricular myocytes indicated that CK-136 increased cellular shortening, with no changes in diastolic Ca²⁺ concentration, systolic Ca²⁺ transient amplitude and time to 75% decline of the Ca²⁺ transient⁷⁶. The researchers did not report on relaxation velocity. A subsequent study showed that TA1, a CK-136 derivative, improved cardiac contractile performance (developed pressure) in healthy rats but did not alter heart rate, diastolic pressure or energetic expenditure⁷⁷. The researchers suggest that by directly targeting the sarcomere, TA1 can increase Ca²⁺ sensitivity and contractility without off-target effects such as increased cAMP signalling and Ca²⁺ cycling, which exacerbate the energetic deficits reported with other inotropes (such as dobutamine). A phase I clinical study was completed by Amgen, but no results have been made available (Table 1).

Myosin activators

As an alternative to sensitizing the regulatory proteins to Ca²⁺, increased cardiac contractility can also be achieved by directly activating cardiac myosin (Fig. 3). In 2010, omecamtiv mecarbil was discovered by high-throughput screening as the first selective, small-molecule activator of cardiac myosin heavy chain-α and cardiac myosin heavy chain-β⁷⁸, but not of other striated-muscle myosins¹³. Omecamtiv mecarbil was initially described as a direct myosin activator that accelerates the actin-activated rate of phosphate release and structurally primes the myosin molecule for interaction with actin^13,79-82. Paradoxically, omecamtiv mecarbil actually inhibits the working stroke of myosin⁸² and is predicted to stabilize the SRX state of myosin⁸³. Although structural studies suggest that omecamtiv mecarbil does not induce the IHM state of myosin²⁵, omecamtiv mecarbil does stabilize the pre-powerstroke state of myosin⁸⁴, which has been associated with the SRX state⁴⁶. Therefore, omecamtiv mecarbil might stabilize a structurally distinct SRX state, although the direct effects of omecamtiv mecarbil on the SRX state have not been measured.

Fig. 3 ∣. The molecular mechanisms of myosin modulation by targeted small molecules.

The myosin heavy chain has a primary role in muscle contraction through the formation of force-generating cross-bridges with actin thin filaments. The steps in the chemomechanical ATPase cycle of myosin are depicted in the context of the basic structural changes, energetics and their role in force production. Small molecules have been developed that target cardiac myosin. Aficamten, blebbistatin and mavacamten all decrease the rate of inorganic phosphate (P_i) release from myosin, leading to reduced ATPase activity of myosin and contractile force production^{79-82,108,127,128}. Additionally, mavacamten increases the population of myosin heads in the autoinhibited, super-relaxed (SRX) state^46,47, whereas blebbistatin, and probably also aficamten, exert their inhibitory effects in an SRX-independent mechanism in cardiac muscle^49,129. Danicamtiv and omecamtiv mecarbil both increase the rate of P_i release from myosin, thereby increasing the ATPase activity and contractile force^79-82,97,98. Danicamtiv also increases the number of myosin heads available for actomyosin cross-bridge formation⁹⁷, which contributes to the increase in contractile force⁹⁸. DRX, disordered relaxed state of myosin.

Despite these inhibitor-like molecular mechanisms, the positive inotropic effects of omecamtiv mecarbil have been explained by simulating the cooperative activation of the thin filament⁸². By prolonging the actomyosin attachment of a small population of non-force-generating myosins, omecamtiv mecarbil activates the thin filament to recruit more drug-free myosins for cross-bridge cycles⁸⁵, with the net result of increasing the number of myosin heads in a force-generating state. This mechanism is highly dependent on the concentration of the drug: higher doses slowed relaxation to the point of ischaemia, as also demonstrated in preclinical and clinical studies^13,86-88. Indeed, at higher doses, omecamtiv mecarbil increased myocardial oxygen consumption in a pig model with left ventricular dysfunction⁸⁶, but at a lower dose, no changes in myocardial oxygen consumption were observed in rat and dog models of heart failure^13,87. Additionally, adverse effects were observed at higher concentrations of omecamtiv mecarbil in the first-in-human study, with signs of myocardial ischaemia due to prolonged contraction⁸⁸.

In the COSMIC-HF trial⁸⁹, which assessed the maximum concentration of omecamtiv mecarbil in the plasma as the primary end point, patients with heart failure with reduced ejection fraction had significant improvements in all the secondary end points of left ventricular systolic ejection time, stroke volume and left ventricular end-systolic diameter, as well as a reduced plasma level of N-terminal pro-B-type natriuretic peptide (NT-proBNP). In the follow-up COSMIC-HF study⁹⁰, patients with heart failure with reduced ejection fraction who were treated with omecamtiv mecarbil also had improved right ventricular structure and function compared with the placebo group. Right ventricular dysfunction commonly results from elevated left ventricular pressure in patients with heart failure and is associated with poor clinical outcomes⁹¹. To date, one phase III trial of omecamtiv mecarbil has been completed (the GALACTIC-HF trial⁹²) and another is ongoing (the METEORIC-HF trial⁹³) (Table 1). Although the results of the GALACTIC-HF trial⁹⁴ showed an absolute reduction in cardiac events or cardiovascular death of only 2.1% with omecamtiv mecarbil treatment, no evidence of an increased risk of myocardial ischaemic events, ventricular arrhythmias, or cardiovascular or all-cause death was found, confirming the safety of the drug dosage in humans. Omecamtiv mecarbil also seemed to be safe in conjunction with other standard therapies for heart failure, given that all the patients in the GALACTIC-HF study were receiving standard-of-care treatment. Importantly, the GALACTIC-HF study^92,94 indicated that omecamtiv mecarbil was of greater benefit in patients with a left ventricular ejection fraction ≤28%. Additionally, a post-hoc analysis of the GALACTIC-HF data identified a clinically significant reduction in the composite end point of time to first cardiovascular event or cardiovascular death among patients classified with severe heart failure⁹⁵. Therefore, omecamtiv mecarbil might hold promise for patients diagnosed with substantially reduced systolic function and severe heart failure, for whom current treatment options are limited.

A promising novel cardiac myotrope is danicamtiv, formerly known as MYK-491. Danicamtiv also targets cardiac myosin and selectively increases cardiac actomyosin activity by increasing phosphate release rates and myosin head availability^96-98. In vivo, danicamtiv had little effect on left ventricular diastolic stiffness, relaxation and Ca²⁺ homeostasis in an experimental model of heart failure in dogs and in a phase IIa study in patients with heart failure with reduced ejection fraction^96,99. Danicamtiv also improved left atrial volume and function⁹⁹, an effect that warrants further investigation and which could be relevant in patients with atrial fibrillation associated with heart failure with reduced ejection fraction. Omecamtiv mecarbil and danicamtiv were directly compared in human engineered heart tissues from induced pluripotent stem cell-derived cardiomyocytes cultured onto decellularized porcine myocardial slices⁹⁸. Although the two small molecules had similar effects on the engineered tissues, such as a dose- dependent positive inotropy associated with changes in contraction kinetics, danicamtiv seemed to improve systolic contraction at a smaller lusitropic cost than that with omecamtiv mecarbil⁹⁸. Currently, patients are being recruited for an exploratory phase II clinical trial designed to test the use of danicamtiv to treat patients with DCM resulting from variants in either MYH7 or TTN¹⁰⁰ (Table 1).

Sarcomeric contractile inhibitors

In this section, we discuss the efforts being made to develop small molecules to inhibit the thin filament and preclinical and clinical data on the molecular mechanisms of myosin-targeted inhibitors.

Thin filament inhibitors

A possible strategy to counteract hypercontractility is to desensitize thin filaments to Ca²⁺. In cardiomyocytes, myofilament Ca²⁺ sensitivity is reduced by β-adrenergic signalling via protein kinase A phosphorylation of troponin I at serine 23 and serine 24^101,102. This mechanism is important to attenuate the myofilament force response to the overall positive inotropic effect of β-adrenergic signalling. β-Blockers have been used for five decades to treat a variety of cardiac diseases, including obstructive HCM¹⁰³.

Furthermore, molecules developed to inhibit calmodulin^104-106 could theoretically be used to reduce myofilament Ca²⁺ sensitivity owing to the structural homology between calmodulin and cardiac troponin C. However, different pharmacological classes of calmodulin antagonist have various effects on cardiac contractility; some increase contraction (such as bepridil), whereas others reduce contraction (such as W7)¹⁰⁵. Currently, no molecules that selectively reduce Ca²⁺ sensitivity by acting directly on the thin filaments are clinically available, but some preclinical efforts are being made¹⁰⁷.

Myosin inhibitors

Mavacamten (previously known as MYK-461) is a small molecule identified by MyoKardia (acquired by Bristol Myers Squibb) in a screening, with the objective to reduce sarcomere contractility in HCM¹⁴. This small molecule binds to myosin and decreases the rate of inorganic phosphate release, stabilizing the SRX conformation of myosin^46,47,108 (Fig. 3). Interestingly, in skeletal muscle, mavacamten increased the SRX population of myosin molecules that were not located in the MyBP-C zone³⁵, suggesting a spatially distinct effect of this pharmacological intervention in the sarcomere.

Mavacamten reduced myosin ATPase activity in mouse and bovine cardiac myofibrils, and reduced fractional shortening with no changes on Ca²⁺ transients in rat cardiomyocytes¹⁴ In vivo preclinical studies on mouse, cat, dog and minipig models of HCM showed that mavacamten treatment reduced fractional shortening, cardiac fibrosis and cardiomyocyte disarray^{14,46,109,110}. Importantly, mavacamten has been shown to reduce diastolic stiffness, probably owing to stabilization of the SRX state of myosin, while still maintaining the length-dependent increase in Ca²⁺ sensitivity (the Frank–Starling relationship) in response to sarcomere stretch^111,112. Mavacamten has also been shown to increase the rate of ADP release from myosin, ATP attachment to myosin and the kinetics of cross-bridge detachment in mouse cardiac strips¹¹² and human tissues^111,113, all of which contribute to reduced diastolic stiffness.

Preclinical studies on the use of a myosin inhibitor in the setting of non-myosin-based disease are very promising. For example, mavacamten can reverse the increase in active myosin motors caused by genetic variants in either MYH7 or MYBPC3^110,114, despite the discrete localization of mavacamten outside of the C-zone of the sarcomere³⁵. Mavacamten also decreased maximum contractile force and Ca²⁺ sensitivity in a mouse model of HCM caused by a point mutation in Myl2, encoding the ventricular/cardiac muscle isoform of the myosin regulatory light chain 2¹¹³. Mavacamten also partially reversed the increased myofilament Ca²⁺ sensitivity and dysregulation of Ca²⁺ flux in cardiomyocyte models of HCM-causing variants in TNNI3 and TNNT2, encoding cardiac troponin I and troponin T, respectively¹¹⁵.

In the phase II MAVERICK-HCM trial¹¹⁶, mavacamten was well tolerated in patients with non-obstructive HCM and was associated with significant reductions in plasma NT-proBNP and cardiac troponin I levels compared with placebo. Patients with obstructive HCM also had a reduction in symptoms with mavacamten in the open-label, phase II PIONEER-HCM trial¹¹⁷ and the phase III EXPLORER-HCM trial¹¹⁸. In the EXPLORER-HCM trial¹¹⁹, the primary end point (a ≥1.5 ml/kg/min increase in peak oxygen consumption and an improvement by at least one NYHA class or a ≥3.0 ml/kg/min increase in peak oxygen consumption with no worsening in NYHA class) occurred in 37% of patients in the mavacamten group and 17% of patients in the placebo group (P < 0.0005) (Table 1). One of the more promising aspects of this trial was that patients with known variants in genes encoding sarcomeric proteins as well as those without known variants responded well to mavacamten¹¹⁹. A subset of patients in the phase III EXPLORER-HCM trial also underwent cardiac MRI, showing that mavacamten treatment significantly reduced left ventricular mass, left atrial mass index and left ventricular wall thickness, with no changes in cardiac fibrosis compared with placebo¹²⁰. These structural changes correlated with significant reductions in the plasma levels of NT-proBNP and high-sensitivity cardiac troponin I¹²⁰. A long-term extension study of the safety of mavacamten (the MAVA-LTE trial¹²¹) is ongoing in patients with HCM who have completed the MAVERICK-HCM or EXPLORER-HCM trials. Preliminary, interim results on the MAVERICK-HCM cohort indicate that long-term treatment with mavacamten (data presented through to week 48) is generally well tolerated, with only mild or moderate adverse events¹²². These patients had a significant reduction in plasma NT-proBNP level (by 67% at week 48) and improved left ventricular relaxation and diastolic function¹²². A potential disadvantage of mavacamten is its slow clearance and half-life of ~9 days¹²³, because these parameters increase the risk of drug toxicity, interactions with other drugs and longer times to achieve steady-state concentrations in the plasma.

Other small molecules (MYK-581 and MYK-224) with similarities to MYK-461 (that is, mavacamten) have been described. Preliminary preclinical data on MYK-581 have been presented at conferences^124,125, and a phase I study of MYK-224 is ongoing according to the Australian New Zealand Clinical Trials Registry¹²⁶.

A preclinical comparison of mavacamten and blebbistatin (a non-specific inhibitor of myosin II) revealed that these two small molecules work via different mechanisms. Blebbistatin has been shown to bind near to the phosphate-binding site in the myosin motor, inhibiting phosphate release and slowing myosin ATPase activity^127,128 (Fig. 3). Unlike mavacamten, blebbistatin has minimal effect on the myosin SRX population in skeletal or cardiac myosin⁸³. Instead, blebbistatin is likely to stabilize a relaxed state of myosin that is structurally distinct from the IHM²⁴, but that still results in ultra-slow myosin ATPase activity.

Another myosin inhibitor has been developed by Cytokinetics called aficamten (also known as CK-274 or CK-3773274). Aficamten was shown to reduce myosin ATPase activity in bovine cardiac myofibrils and to reduce contractility in rat primary cardiomyocytes¹²⁹. Although the effects of aficamten on the SRX state of myosin have not been reported, aficamten has been shown to bind to the ATP-binding pocket of myosin¹²⁹, in a similar manner to that of blebbistatin. Therefore, aficamten is likely to stabilize a myosin off-state that is structurally distinct from the closed state induced by mavacamten⁸³. Despite the predicted unique molecular mechanisms of inhibition, preliminary data indicate that aficamten reduces cardiac contractility in vivo in healthy rats and beagle dogs, similar to the effects of mavacamten^130,131 (Fig. 3), as well as in a transgenic mouse model of HCM carrying the R403Q variant in myosin¹³². A detailed assessment of aficamten reported faster pharmacokinetics than those of mavacamten, with a predicted half-life in humans of 2.8 days, an actual half-life of 3.4 days and no substantial effect on cytochrome P450^129,133, suggesting safety and tolerability of the drug for therapeutic interventions. Currently, the phase II REDWOOD-HCM trial¹³⁴ is recruiting patients with outflow-tract obstruction HCM to assess the incidence of adverse events during dosing of aficamten (Table 1). Preliminary results presented at a conference and in a press release from the company¹³⁵ report that aficamten treatment significantly reduced resting left ventricular outflow tract gradients and the plasma NT-proBNP level compared with placebo. Aficamten was also generally well tolerated, with no serious adverse events reported¹³⁵. A phase III trial of aficamten is expected to start soon.

Conclusions

In the past 20 years, major scientific advances have been made in understanding the pathophysiology and genetics of cardiovascular disease. However, drug development for these diseases has been comparatively stagnant, in large part owing to the high costs and risks associated with the clinical management of cardiovascular disease¹³⁶. Although druggable targets have been identified in the cardiac sarcomere for direct modulation of contractile function in these diseases, many unknowns remain in the field of sarcomere pharmacology. One major question, given the vast number of variants in such a large number of genes and several mechanisms of pathogenesis, is whether one or a small number of small molecules targeting sarcomeric proteins will be effective in a heterogeneous spectrum of diseases. Of note, mavacamten was effective in patients with HCM regardless of whether they had an identified variant in a gene encoding a sarcomeric protein¹¹⁹. Consistent with these results is a report that the proteomes of patients with HCM and variants in different genes encoding sarcomeric proteins converge downstream of the primary genetic variant¹³⁷. Clinical data from patients receiving aficamten treatment will help to refine the question of whether mechanistically distinct small molecules, such as the myosin inhibitors aficamten and mavacamten, have more or less benefit in particular subgroups of patients with HCM.

A phase II clinical trial of danicamtiv has recruited patients with DCM and genetic variants in either MYH7 or TTN. The preclinical development of compounds targeting MyBP-C¹³⁸ and research on their efficacy in patients with MYBPC3 or other gene variants will continue to explore the interactions between sarcomeric proteins and their direct versus indirect modulation in clinical management. Although the developmental and tissue-specific expression patterns of sarcomeric proteins complicate the development of compounds targeted to the sarcomere, the studies discussed in this Review highlight the benefit of directly modulating contractile function in cardiac muscle, supporting myofilament-targeted compounds as the future of muscle-based therapeutics.

Key points.

• Variants in genes encoding sarcomeric proteins are a leading cause of cardiomyopathies, characterized by protein-specific molecular mechanisms of disease.

• Sarcomeric proteins can be targeted by small molecules to directly modulate contractile function in cardiac muscle.

• Small molecules that are targeted to the myosin heavy chain modulate enzymatic activity and/or availability of the motor, leading to an increase or decrease in force production.

• Small molecules that target the myosin heavy chain, such as mavacamten and aficamten, can act via distinct molecular mechanisms that lead to altered myosin function.

• Sarcomere pharmacology suggests that small molecules targeting a specific sarcomeric protein will be effective even in patients with a causal variant in a gene encoding another sarcomeric protein.