Novel Therapies for Prevention and Early Treatment of Cardiomyopathies.

Heritable cardiomyopathies are a class of heart diseases caused by variations in a number of genetic loci. Genetic variants on one allele lead to either a degraded protein, which causes a haploinsufficiency of that protein, or a nonfunctioning protein that subverts the molecular system within which the protein works. Over years, both of these mechanisms eventually lead to diseased heart tissue and symptoms of a failing heart. Most cardiomyopathy treatments repurpose heart failure drugs to manage these symptoms and avoid adverse outcomes. There are few therapies that correct the underlying pathogenic genetic or molecular mechanism. This review will reflect on this unmet clinical need in genetic cardiomyopathies and consider a variety of therapies that address the mechanism of disease rather than patient symptoms. These therapies are genetic, targeting a defective gene or transcript, or ameliorating a genetic insufficiency. However, there are also a number of small molecules under exploration that modulate downstream faulty protein products affected in cardiomyopathies.

C ardiomyopathies are primary heart disorders that occur in the absence of underlying causes such as coronary artery disease, hypertension, and valvular or congenital heart disease. 1 Based on the predominant clinical abnormalities in cardiac structure and function, cardiomyopathies are classified into 3 major subtypes: hypertrophic, dilated, and arrhythmogenic. Discoveries made during the past several decades have defined the precise genetic pathogenesis in many patients with these disorders, which have propelled insights into the molecular mechanisms by which pathogenic variants cause cardiomyopathies. These advances raise the prospect for new treatments that directly target gene variants or the proximal downstream pathways that mediate disease.
Multiple genetic approaches have the potential to directly attenuate the deleterious effects of pathogenic cardiomyopathy variants (Table). These include targeted correction of the specific causal variant-an approach that would ameliorate any pathogenic variant. For the subset of cardiomyopathies that result from variants that cause insufficient protein levels, gene delivery systems to increase protein could be effective. When cardiomyopathies arise from dominant-negative missense proteins, emerging strategies aim to modify or silence the mutant transcript. A range of alternative molecular strategies have been developed to address these approaches because of complexities that include heterogeneity of human pathogenic variants, gene and transcript sizes, and other determinants that impact feasibility (Table). Concurrent with these genetic strategies, advances have also been made in the development of small molecules that repair or diminish the biophysical consequences of mutant proteins. Progress in these traditional medicinal approaches has led to several ongoing clinical cardiomyopathy trials. This review will discuss opportunities and challenges for advanced precision therapies, including genetic and small-molecule strategies, to improve outcomes and unmet clinical needs in genetic cardiomyopathies.

Hypertrophic Cardiomyopathy
Hypertrophic cardiomyopathy (HCM) is a common disorder characterized by left ventricular wall thickening in the absence of other cardiac or systemic disease-a phenotype that occurs in ≈1 in 500 individuals. [2][3][4] Clinical hallmarks of HCM include findings from echocardiography or cardiac magnetic resonance imaging, such as asymmetrical hypertrophic remodeling with greater septal involvement, dynamic left ventricular outflow tract obstruction, nondilated ventricular chambers with mitral valve abnormalities, and left atrial enlargement. 5 At the cellular level, HCM manifests as cellular hypertrophy, cardiomyocyte disarray, as well as accumulative focal or interstitial fibrosis. The prototypic functional changes in HCM are hyperdynamic contraction, progressive diastolic dysfunction, and increased energy consumption. 6 These phenotypes accompany the development of exertional dyspnea, angina, and palpitations and are presumed to promote many adverse outcomes, including atrial fibrillation with risk of thromboembolic events, progressive development of heart failure, and occasionally, sudden cardiac death.
Current clinical HCM therapies focus on symptom management and prevention of sudden cardiac death. Symptom management includes the use of pharmacological agents to improve cardiac relaxation and energetics as well as septal reduction strategies to relieve outflow tract obstruction. Additionally, automatic implantable cardiac defibrillators and antiarrhythmic drugs are used in the prevention of sudden cardiac death. Despite these treatment options, meta-analyses of almost 5000 patients at HCM centers of excellence (inclusive of 25 000 longitudinal patient-years) showed that atrial fibrillation and heart failure continued to emerge in over 20% of patients with HCM, events that increased with disease duration. 7 By contrast, the frequency of ventricular arrhythmias (6%) and sudden cardiac death (4%) was low. An important inference from these data is that current medical therapies are unable to prevent disease progression, which promotes atrial fibrillation and heart failure.
Dominant pathogenic variants in the thick and thin filament sarcomere proteins are the predominant genetic cause of HCM. 8 Although over 1400 different pathogenic variants have been reported, the majority (70%-80%) alter the genes that encode MYH7 and MYBPC3. Less commonly, pathogenic variants are identified in genes encoding thin filament proteins TNNI3 (troponin I) or TNNT2 (troponin T) and rarely in genes TMP1 (tropomyosin), ACTC1 (actin), and TNNC1 (troponin C), or other sarcomere-associated proteins. 9,10 Most pathogenic variants that cause HCM are heterozygous missense variants, which substitute a single amino acid in 1 of 2 gene copies that encode a sarcomere protein. Mutant proteins are incorporated into the sarcomere, where they perturb performance and trigger cardiac remodeling. Unlike most HCM genes, pathogenic variants in MYBPC3 usually encode premature termination signals and result in either unstable transcripts or truncated cMyBP-C (cardiac myosin-binding protein-C) peptides that lack myosin-binding and titin-binding sites. 11 These loss-of-function (LoF) transcripts and proteins typically undergo nonsense-mediated decay and thereby reduce the amount of cMyBP-C protein in sarcomeres below normal levels (haploinsufficiency). Because clinical phenotypes are similar between HCM patients with pathogenic MYH7 missense variants or MYBPC3 truncating variants, both are expected to evoke comparable biophysical and biochemical abnormalities in sarcomeres.
Dominant transmission of HCM in affected families indicates that first-degree relatives have a 50% risk for carrying the mutant allele. Moreover, longitudinal clinical assessments of carriers of pathogenic variants demonstrate age-related penetrance of hypertrophy that typically manifests near the age of puberty. However, detailed cardiac assessments of preclinical individuals, those with pathogenic variants but without hypertrophy, demonstrate hypercontractility, diastolic dysfunction, and increased energy consumption. [12][13][14] The presence of these abnormalities in preclinical mutation carriers indicates that the pathophysiology of HCM precedes the onset of hypertrophy. As such, therapeutic interventions may need to be administered early to prevent cardiac remodeling or else be capable of reversing overt disease.

Genetic Approaches Targeting HCM Variants
Because HCM is caused by hundreds of different dominant acting missense or LoF variants in a sarcomere protein, the most direct approach to prevent disease development would be to correct each pathogenic variant before clinical manifestation. A proof-of-concept study aimed to perform prezygotic correction by ex vivo DNA manipulation of sperm that carried a 4-base pair pathogenic MYPBC3 deletion and an egg with a normal MYPBC3 allele. 15 This approach used CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/clustered regularly interspaced short palindromic repeat-associated 9) technology where CRISPR components were injected to induce sequence-specific breaks in the paternal allele, which were repaired by homology-directed repair using oligonucleotides that provided a homology-directed repair template to correct the MYBPC3 variant ( Figure 1). DNA breaks can also be repaired by nonhomologous end joining with insertion or deletion of sequences. When reagents were delivered 18 hours post-fertilization, 33% of embryos remained heterozygous for the MYPBC3 pathogenic variant or were mosaics, containing an admixture of corrected and mutant cells. In a modified approach, CRISPR reagents were simultaneously coinjected with sperm. While this resulted in 72% of embryos with normal biallelic MYPBC3 sequences and no evidence for mosaicism, 28% retained the MYPBC3 pathogenic variant, alone or accompanied by an insertion or deletion. Unexpectedly, restoration of the normal MYBPC3 sequences occurred with maternal DNA acting as the homology-directed repair template, not the exogenously delivered oligonucleotide template-a finding that may imply intrinsic embryonic repair mechanisms that remain to be understood. A considerable technical concern associated with direct genomic manipulation in early embryogenesis is the potential for off-target editing that could cause mutagenesis and subsequent damaging variants. Whole genome sequencing of embryos from the study described above identified multiple new insertions and deletions that occurred within poly-A or poly-GT repeats. Further studies are needed to determine whether these variants reflect errors in sequencing or off-target events with potential adverse functional consequences. This approach also raises substantive social, legal, and ethical considerations regarding the use of embryos for research or heritable germline editing. Moreover, clinical gene-based diagnosis of preimplantation embryos can prevent HCM transmission. As such, these studies demonstrate feasibility for ex vivo gene corrective strategies in HCM, but clinical application of this approach remains a distant reality.
Other preclinical genetic strategies to treat HCM selectively target in vivo LoF variants or damaging missense variants in mouse models. These approaches deliver therapeutic agents using viral vectors. Lentivirus, which accommodates large-sized inserts (≈10 Kb), integrates into the host genome at unpredictable and potentially deleterious sites. Adenoassociated virus serotype 9 (AAV9) accommodates inserts <5 Kb and has cardiotropic activities. While viral-mediated genetic therapies are in development for several human disorders, 16 preexisting antibodies or the induction of immune responses to viral capsids reduces the efficiency of reinduction by the same viral construct. In addition, preclinical studies demonstrate nonuniform transduction of cardiomyocytes, nonregulated expression of the viral cargo, and progressive decline of viral titers within transduced cells.
Gene replacement therapies have aimed to increase cMyBP-C protein in mouse models harboring Mybpc3 truncating variants that lead to haploinsufficiency ( Figure 2). Lentiviral delivery of full-length mouse Mypbc3 cDNA to the myocardium of homozygous Mybpc3-null mice restored normal protein content and cross-bridge kinetics in isolated heart tissues, and in vivo assessments revealed improved ventricular function. 17,18 A related approach used AAV9 carrying the mouse Mybpc3-cDNA that was expressed under the control of the troponin promoter. 19 A single injection into neonatal homozygous LoF Mybpc3-mutant mice both restored normal cMyBP-C protein levels and suppressed low-level production of defective mRNA species from the mutant alleles. This strategy prevented the development of HCM for ≈6 months, after which suppression was lost, presumably because of decline in AAV9 titers. Parallel studies used in vitro transduction of MYBPC3-cDNA into primordial human cardiomyocytes at day 12 after differentiation from mutant embryonic stem cells with a heterozygous MYBPC3 LoF variant. Within a week of gene delivery, protein levels were normalized, sarcomere organization and calcium impulse propagation were improved, and cellular hypertrophy was prevented. 20 Two alternative approaches have emerged to correct endogenous mutant Mybpc3 transcripts in mice. Unlike replacement strategies that harness active promoters to induce high levels of expression, correction of endogenous mutant transcripts could maintain physiological cMyBP-C protein levels. Mice carrying a homozygous Mybpc3 G>A transition in the last nucleotide of exon 6, the −1 base of a canonical splice donor sequence, express multiple mutant Mybpc3 transcripts and develop HCM. 21 One of these mutant transcripts excludes exons 5 to 6 but otherwise retains the normal coding frame. 22 Delivery of antisense oligoribonucleotides that mask splice enhancers in Mypbc3 exons 5 and 6 resulted in increased expression of the shorter but functional protein and transiently prevented hypertrophy and cardiac dysfunction.
A second approach to correct mutant transcripts capitalized on trans-splicing between 2 alleles (Figure 3). Cardiomyocytes derived from human induced pluripotent stem cells (hiPSC-CMs) with an MYBPC3 LoF allele were transduced with virus carrying a construct with a segment of the normal MYBPC3-cDNA fused to strong splice sequences and a binding domain that targets a specific site in the mutant prespliced transcript. Trans-splicing between the endogenous CRISPR/Cas9 indicates clustered regularly interspaced short palindromic repeats/clustered regularly interspaced short palindromic repeat-associated 9; LoF, loss of function; miRNA, micro-RNA; SERCA, sarcoplasmic reticulum calcium-ATPase; and siRNA, small interfering RNA. mutant transcript and WT (wild type) sequences resulted in a fully corrected chimeric MYBPC3 transcript. Although this approach could allow the correction of multiple different pathogenic variants, the low efficiency of trans-splicing represents a significant barrier. However, AAV9-delivered MYBPC3 cDNA into hiPSC-CMs harboring this MYBPC3 variant achieved 80% of WT cMyBP-C protein incorporation. 23 Pathogenic missense variants that cause HCM present additional challenges, as these are stably expressed, incorporated into the sarcomere, and cause dominant-negative effects. Strategies, such as trans-splicing, that increase the proportion of normal transcripts, might effectively dilute mutant sarcomere proteins and thereby reduce disease. An alternative strategy is to selectively silence the mutant transcript using small interfering RNAs (siRNAs), allowing the endogenous WT allele to make up the sarcomeric protein content. This approach is predicated on the recognition that many sarcomere genes that harbor HCM missense variants (eg, MYH7, TNNT2, and TNNT3) are not dosage sensitive, with the exception of MYBPC3, meaning that the expression of 1 allele is sufficient for normal cardiac function. Human genetics supports this conclusion, as broad-based population exome sequencing predicts that LoF variants in many sarcomere genes are tolerated, 26 unlike damaging missense variants that cause HCM.
To silence damaging missense transcripts, AAV9 has been harnessed to deliver siRNAs that target the heterozygous pathogenic myosin variant expressed in mice. The construct that directed allele-specific silencing was under the control of the troponin promoter so as to enable highly efficient transduction of cardiomyocytes. 27 In comparison to untreated mice, allelespecific silencing caused a 25% reduction in mutant transcripts, which was sufficient to prevent the development of hypertrophy, fibrosis, and molecular markers of disease for 6 months. Because the development of specific siRNAs would be needed to address the hundreds of unique missense variants that cause HCM, an alternative approach used allele-specific siRNAs that targeted nearby single-nucleotide polymorphisms rather than the causal variant. Targeting common single-nucleotide polymorphisms that occur in the vicinity of many HCM variants also silences the mutant allele and could eliminate the need to develop tailored siRNAs for each pathogenic variant.

Genetic Amelioration of Signals Evoked by HCM Mutations
Abnormal calcium homeostasis occurs in many pathogenic HCM variants, including those in thick and thin filament proteins. 28 In addition to critical roles in excitation-contraction coupling, calcium serves as a universal signaling molecule that likely contributes to hypertrophic remodeling. Cardiac SERCA2a (sarcoplasmic reticulum calcium-ATPase 2a) promotes rapid calcium reuptake into the sarcoplasmic reticulum, which is inhibited by phospholamban and disinhibited by phosphorylation of phospholamban. Increasing the ratio of SERCA2a:PLN (phospholamban), by suppressing PLN or overexpressing SERCA2a, would shorten relaxation times and improve diastolic dysfunction-a prominent feature of HCM. 29-31 AAV-mediated delivery of the Serca2a gene into newborn transgenic mice that overexpress a human HCM missense variant in α-tropomyosin demonstrated reduced hypertrophy and fibrosis, as well as normalized hemodynamics when compared with untreated mice. 32,33 An alternative approach deleted phospholamban (via crosses with Pln −/− mouse), which eliminated a physiological Serca2a inhibitor so as to increase Serca2a-mediated calcium reuptake. 34 Compound α-tropomyosin/Pln −/− mice had normalized cardiac size, reduced collagen deposition, and improved ventricular function. Unfortunately, strategies that might deplete human PLN expression in humans are unlikely to be therapeutic because heterozygous PLN LoF variants can cause dilated cardiomyopathy (DCM). 35,36 FHLs (four and a half LIM [Lin-1, Isl1, Mec3] domain), Fhl1 and Fhl2, are z-disc proteins that inhibit pathological Cas9-an endonuclease DNA enzyme-is guided by sequence-specific RNA to cleave complementary target DNA. Once double-strand breaks are induced, DNA is repaired and modified either by homology-directed repair (HDR; left side) or by nonhomologous end joining (NHEJ; right side). During HDR, the cell uses an exogenously introduced template to introduce a specific missense mutation, whereas NHEJ generates insertions or deletions during double-strand break repair that often lead to a frameshift and, therefore, produces a loss-of-function variant in the targeted allele. activation of calcineurin 37-39 -a phosphorylase that activates nuclear factor of activated T-cell transcription factor that often promotes hypertrophy. 40 Depletion of Fhl1 increased hypertrophic remodeling in HCM mouse models, 41 whereas viralmediated increased expression of Fhl2 provided protection from hypertrophic remodeling in mouse cardiomyocytes and engineered heart tissues. 42 A related approach has targeted a cardiac-enriched interacting protein encoded by Cefip that opposes Fhl2. SiRNA depletion of Cefip transcript repressed calcineurin activity and hypertrophy. 43 Together, these studies indicate the potential salutary effects of modulating HCM signaling that emanates from z-disc molecules.

Emerging Pharmacological Strategies in HCM
Damaging missense and LoF variants in sarcomere proteins can cause similar and fundamental changes in protein-protein interactions within the sarcomere that drive HCM. MYH7 missense variants and MYBPC3 truncations destabilize the myosin interacting-heads motif, reducing the number of myosins that form a sequestered, energy conserving state of myosin, termed the super relaxed state. This results in a greater proportion of myosin heads in a disordered relaxed state-a state in which myosin can more readily interact with actin and utilize ATP to drive muscle contraction ( Figure 4A). [44][45][46][47] This mechanism accounts for increased contractility and energy consumption, as well as poor relaxation, and implies that strategies to modulate the myosin interacting-heads motif could be effective across many genetic causes of HCM. Moreover, pharmacological strategies that combat common molecular mechanisms underlying HCM (or other cardiomyopathies) provide the opportunity to treat many patients, irrespective of the precise genetic pathogenesis.
MYK-461 is a small-molecule myosin ATPase inhibitor that was initially developed to reduce contractility in cardiomyocytes. 48 Oral treatment of preclinical mice that carried 1 of 3 different pathogenic myosin missense variants with MYK-461 reduced the development of ventricular hypertrophy, cardiomyocyte disarray, and myocardial fibrosis and normalized the expression of profibrotic and mitochondrial genes. More recent studies found that MYK-461 affected the myosin interactingheads motif and could sequester myosin heads from their disordered relaxed state to super relaxed state. 45 Proof-of-concept studies have demonstrated that Mybpc3 mouse models of HCM, hiPSC-CMs, and human heart tissues with MYBPC3 LoF variants destabilize the interacting-heads motif. 47,49,50 Treatment with MYK-461 rebalanced super relaxed state:disordered relaxed state myosin ratios and corrected HCM phenotypes of cellular hypercontractility and poor relaxation ( Figure 4A). 47 Figure 3. Strategies to correct aberrant transcripts. Schematic representation of (A) trans-splicing-a reaction that occurs between a targeted endogenous mutant pre-mRNA (orange) and a therapeutic pre-trans-splicing molecule (blue) and that results in a processed mRNA transcript devoid of the exon containing the pathogenic variant. B, Exon skipping removes the exon containing a pathogenic variant, while maintaining the correct reading frame. A treatment for Duchenne muscular dystrophy uses an antisense oligonucleotide to mask exon 51, resulting in the splicing that excludes exon 51. 24 C, Micro-RNA (miRNA) and (D) small interfering RNA (siRNA) are short RNA molecules that exert gene-silencing effects on RNA transcripts. miRNA molecules are nonspecific and can have many targets, whereas siRNAs are sequence-specific and can be designed to only target a selected mRNA. miRNA binds to mRNA transcripts through partial complementary base pairing, and gene silencing occurs via translational repression or transcript degradation/cleavage. Conversely, a siRNA is fully complementary to the coding region of its target mRNA, and its binding ultimately results in cleavage of the transcript. 25 An initial open-label trial of MYK-461 (mavacamten) given to 11 HCM patients with outflow tract obstruction reported significant reduction in postexercise gradients, improved peak Vo 2 , New York Heart Association functional class, dyspnea scores, and N-terminal pro-BNP (brain natriuretic peptide) levels at 12 weeks of treatment. 51 A multicenter, randomized, double-blind trial in 220 obstructive HCM patients is underway. 52 HCM also impairs myocardial metabolism and increased oxidative stress 53,54 because of excessive ATP usage or inadequate fuel delivery from microvascular dysfunction and outflow tract obstruction. 55 An early attempt to correct the HCM myocardial energy deficit was perhexaline-a carnitine palmitoyl-acyltransferase inhibitor that shifts metabolism from fatty acid oxidation to glycolysis. Follow-up at 4 to 6 months in 24 patients with HCM showed improved peak oxygen consumption during exercise, New York Heart Association class, diastolic function, and myocardial ratio of phosphocreatine:ATP. 56 However, a proposed larger multicenter randomized clinical trial was abandoned because of lack of efficacy. 57 The antioxidant glutathione precursor N-acetylcysteine (NAC) has been studied as a possible therapeutic agent for HCM to reduce myocardial oxidative stress and abate downstream kinases that activate fibrosis. NAC treatment of transgenic mice with pathogenic HCM variants in cardiac troponin T reduced cardiac hypertrophy and fibrosis, 58 and transgenic mice with α-tropomyosin variants showed improved diastolic function. 59 Remarkably, NAC treatment of overt disease in transgenic rabbits that express a human pathogenic MYH7 missense variant led to regression of hypertrophy and fibrosis and reduced ventricular arrhythmogenicity. 60 However, a recent placebo-controlled single-center human trial of high dose (≤2.4 g/d) NAC in HCM (HALT-HCM: Hypertrophy Regression With N-Acetylcysteine in Hypertrophic Cardiomyopathy) resulted in 6 adverse events in 29 patients receiving active compound that were not directly attributed to NAC. Although the sample size was small, 12 months of treatment showed small effects on hypertrophy or fibrosis. 61,62 Dilated Cardiomyopathy DCM is characterized by left ventricular chamber enlargement (end-diastolic diameter >117% normal) and diminished systolic performance (fractional shortening <25% or ejection fraction <45%). 63 DCM, the most common indication for cardiac transplantation, occurs in the setting of an underlying cardiovascular and systemic disorders, including diabetes mellitus, hypertension, and atherosclerosis. DCM also arises secondary to exposures, including cancer therapies and infection. Increasingly, DCM is recognized as a primary genetic disorder that occurs without or accompanies predisposing factors.
Management strategies address underlying conditions but otherwise are directed toward symptom improvement, preservation of myocardial function, and reducing morbidity and premature mortality. Traditional therapeutic approaches focus on β-blockers and ACE (angiotensin-converting enzyme) inhibitors or ARBs (angiotensin II receptor blockers) to improve cardiac systolic function, diuretics for congestion, and management of arrhythmias. An important addition to this armamentarium is LCZ696, a medication that combines sacubitril, an inhibitor of the neutral endopeptidase neprilysin, with the ARB valsartan. 64 Sacubitril increases serum concentrations of natriuretic peptides, bradykinin, and adrenomedullin that can limit the effects of excessive neurohormonal activation that accompanies heart failure. The PARADIGM-HF trial (Prospective Comparison of ARNI [Angiotensin Receptor-Neprilysin Inhibitor] with ACEI [Angiotensin-Converting-Enzyme Inhibitor] to Determine Impact on Global Mortality and Morbidity in Heart Failure) showed superiority of LCZ696 to enalapril in reducing heart failure hospitalizations and risk of death. 64 Cardiac synchronization therapy is used in selected patients who meet guideline criteria, whereas implanted automatic cardiac defibrillators are appropriate when the ejection fraction falls below 35%, because these patients are at greater risk of arrhythmias and sudden cardiac death. 65 When these approaches are ineffective in preventing heart failure development, and progression, patients may receive advanced interventions including ventricular assist devices and cardiac transplantation. The identification of the genetic basis for DCM provides opportunity to develop earlier and more precise treatments that may lessen these adverse outcomes.
Genetic DCM was traditionally recognized by familial clustering, defined as >2 affected relatives or a first-degree relative with unexplained sudden cardiac death before 35 years of age. 66 Clinical estimates suggest the prevalence of familial DCM is 30% to 50% among patients with DCM. 66,67 Because pathogenic variants are identified in ≈40% of familial DCM patients, 68 the genetic architecture of DCM remains incompletely understood. Most pathogenic DCM variants are expressed as dominant traits with variable clinical penetrance.
Pathogenic DCM variants have been identified in over 40 genes that function in many different processes. 66,69 DCM variants mutate genes that encode sarcomere proteins and molecules involved in calcium cycling, sarcomere-associated proteins involved in mechanosensing, and molecules involved in force transmission. Others impair the protection of the nuclear envelope from biomechanical forces, damage heat shock chaperones, injure mitochondria, or perturb other critical subcellular processes. Because some DCM genes may be expressed only in the heart or in other cell lineages, some pathogenic variants produce cardiac-specific disease, whereas others are associated with additional clinical phenotypes.
Damaging variants in titin, encoded by the gene TTN, are the most common cause of DCM. TTN encodes the largest known protein, comprising 35 000 amino acids. The pathogenicity of TTN missense variants is uncertain, and only titin truncating variants (TTNtvs) definitively cause DCM. Although TTNtvs are distributed throughout the gene, those located in exons that undergo alternative splicing are eliminated from the mature transcript and protein. The percent spliced in quantifies the cardiac expression of each TTN exon. TTNtvs that reside in exons with percent spliced in >90% have greater pathogenicity than those encoded in exons with low percent spliced in. 70,71 TTNtvs occur in 25% of familial cases, 72 15% of sporadic cases, 71 and in 10% of DCM that emerges in the context of pregnancy, 73 alcoholic cardiomyopathy, and cancer therapy-induced cardiomyopathy. 74,75 Damaging missense and LoF variants in the nuclear membrane protein gene LMNA (lamin A/C) are the second most common genetic cause of DCM. These usually cause cardiomyopathy in association with progressive electrophysiological deficits. Unlike many DCM genes, LMNA is ubiquitously expressed in all cells, and pathogenic variants can have pleiotropic extracardiac effects, including neuromuscular phenotypes.
The multiplicity of DCM genes provides considerable challenges for developing precision therapies-hundreds of distinct pathogenic variants in many, and often large genes trigger a wide variety of pathophysiologic mechanisms.

Genetic Approaches Targeting DCM Variants
While no germline correction of embryos carrying DCM variants has been reported, patient-derived hiPSC-CMs carrying a pathogenic phospholamban variant (PLN-R14del) were corrected using transcription activator-affected nucleases in combination with homologous recombination directed repair. This approach established biallelic expression of normal PLN transcripts. 76 An alternative to gene correction has been viral-mediated gene supplementation to boost the expression of dosage-sensitive DCM genes ( Figure 2). DCM that rapidly progresses to heart failure develops in a hamster model harboring a deletion in the Sgcd that encodes a component of dystrophin-associated glycoprotein complex. 77 AAV-mediated transfection of the 4 Kb Sgcd cDNA restored the expression of all sarcoglycans and improved but did not fully normalize contractile function. 78 Correction of deficient protein levels is challenging in DCM because of the large size of many of the disease-causing genes and the limitation of current vectors to carry large gene or transcript cargos. For example, the TTN gene encodes 364 exons across 294 Kb of genomic DNA that are spliced into a ≈110 Kb transcript-sizes that vastly exceed the cargo capacity for viruses that are readily transduced into cardiomyocytes (Table). An alternative approach to compensate for TTN haploinsufficiency capitalizes on a conserved internal promoter known as Cronos that is active in human heart tissues and can be used to create a shorter TTN isoform. 79 Analyses in zebra fish showed that activation of this internal promoter partially rescued phenotypes caused by N-terminal truncations encoded upstream of the Cronos promoter. These observations raise the possibility that endogenous Cronos activity might account for incomplete penetrance and more mild cardiac phenotypes in some patients with TTNtv. 80 Moreover, exogenous delivery of molecules that activate the Cronos promoter and increase the expression of a shorter but possibly functional TTN isoform in human hearts could have therapeutic benefits.
Another approach to increase gene expression is to manipulate the processing of mutant RNAs, so as to minimize the pathogenicity of the encoded protein. Pathogenic variants that result in frameshifts often produce a premature stop codon, which can induce nonsense-mediated decay of mutant mRNA, and subsequently protein haploinsufficiency. 81 Inactivation of the molecules participating in nonsense-mediated decay can stabilize mutant transcripts and promote translational readthrough, in which a noncognate transfer RNA inserts an amino acid at the termination codon ( Figure 5). High concentrations of aminoglycosides and small molecules such as PTC124 studied in high-throughput assays showed variable efficacy of ribosomal read-through of premature stop codons but not the normal termination codons. 82 This strategy was also used to mitigate the consequences of 3 different LMNA nonsense and frameshift variants in patient-derived hiPSC-CMs. Administration of PTC124 enabled protein translation in only 1 cell line, which showed reduced cardiomyocyte apoptosis and improved excitation-contraction coupling. However, PTC124 treatment was ineffective in 2 other lines because of flanking codons that inhibited read-through. 83 The apparent codon selectivity of PTC12 has limited extensive development of this approach.
Antisense-mediated exon skipping and trans-splicing ( Figure 3) are also under exploration to remove mutant exons with frameshift or nonsense variants in DCM genes. These approaches resultant in-frame transcripts that encode a shorter peptide that must retain sufficient biological activity to ameliorate disease. This strategy was used to eliminate a TTNtv in exon 326 within a patient-derived hiPSC-CMs. Only 2 of 4 antisense oligonucleotides developed to mediate exon skipping were successful and appeared most efficacious when both were delivered together. 85 The considerable diversity and distribution of TTNtv across many exons may hinder the widespread use of this approach.
Trans-splicing was studied in Lmna mutant mice with a homozygous amino-terminus single base pair deletion in exon 1, which causes profound systemic and cardiac disease with early lethality. 86 AAV-mediated transduction of an Lmna construct with 5 amino-terminal exons, a 5′ splice donor sequence, and binding domain targeting intron sequences enabled trans-splicing in skeletal and heart muscle, however, at insufficient levels to alter phenotype or premature death. 87

Genetic Amelioration of Signals Evoked by DCM Mutations
Micro-RNAs (miRNAs) bind 3′ untranslated sequences in target mRNAs, promoting degradation of transcripts and translational repression. miRNAs can dynamically modulate contractile proteins, calcium handling, and metabolic signals in response to health, disease, and interventions, 88 making them biomarkers of disease and potentially therapeutic molecules. Mice with homozygous LoF variants in the Sgcb develop severe and progressive DCM because of substantial cardiomyocyte apoptosis. Molecular profiling of mutant hearts identified reduced expression of miRNA-669a, which targets MyoD transcripts. MyoD-a master regulator of skeletal myogenesis-suppresses cell proliferation and promotes differentiation. 89,90 AAV-mediated sustained delivery of miRNA-669a for 18 months decreased cardiac MyoD expression, enhanced cardiac structure and function, and prolonged survival without evidence of adverse effects on skeletal muscle. miR-208b is increased in human DCM patients and in mice with a heterozygous Ttn-null allele were stressed with angiotensin II infusion. 91 Treatment with antisense oligonucleotides directed against miRNA 208b prevented ventricular dilatation, fibrosis, and normalization of stress-responsive myosin isoforms. While the precise mechanisms by which miRNA 208b inhibition improved cardiac phenotypes in these mice are uncertain, miRNA 208b also targets Tharp 1 (thyroid receptor coregulator) and myostatin. 92 Altered expression levels of these proteins might contribute to myosin isoform switching and improve contractile function.

Emerging Pharmacological Strategies in DCM
Small molecules can improve sarcomere performance by increasing troponin affinity for calcium (calcium sensitizers) or by directly binding and modulating actin-myosin interactions. 93 While both approaches increase cross-bridge formation and generate greater sarcomere force, amplifying the finely tuned contractile apparatus will often result in concurrent slowing of cardiac relaxation and increase energy consumption. 94 Calcium sensitizers can also promote inappropriate cardiac depolarization precipitating arrhythmias 95 -a side effect that has frequently impaired the progress of bioactive compounds to the clinic.
Two molecules were recently developed that directly bind myosin, increase actin affinity and cross-bridge formation, and increase sarcomere force production. EMD 57033 96 and CK-1827452 are allosteric modulators that bind in the same region of myosin ( Figure 4B). 97 EMD 57033 bound to myosin increases the rate of ATP binding, hydrolysis, and actin interactions, thereby accelerating the pace of the chemomechanical cycle. In addition, EMD 57033 has chaperone activities that promoted refolding and stabilization of heat-inactivated myosins, as well as restoration of function. EMD 57033 had prohypertrophic effects on unstressed rat neonatal cardiomyocytes. When challenged with heat for 24 hours, cells treated with EMD 57033 were not hypertrophied, maintained normal levels of cardiac myosin proteins, and had no increased expression of atrial natriuretic peptide-a molecular stress marker that was increased in untreated cells. Transcripts encoding both αand β-cardiac myosins were decreased, with larger reduction in β-myosin transcripts. 96 CK-1827452 accelerates the transition of actin-myosin complex from weakly to strongly bound and increases the number of myosin heads engaged with the thin filament. 98,99 Crystal structures show that CK-1827452 bound to myosin primes the lever arm so that more myosin heads are positioned to interact with actin ( Figure 4B). 99 These effects are independent of calcium transients because CK-1827452-treated cardiomyocytes with isoproterenol augment contraction, whereas β-adrenergic inhibition does not diminish contractility. 97 Additional studies indicated that CK-1827452 traps some myosin heads in a weak actin affinity state with slow force development 100 and at high concentrations prolonged cellular relaxation in hiPSC-CMs. 94 A randomized double-blind study showed that CK-1827452 administered to patients with chronic stable symptomatic heart failure increased stroke volume and modestly reduced left ventricular end-diastolic diameter, heart rate, and serum levels of N-terminal brain natriuretic factor. 101 Additional ongoing clinical trials with CK-1827452 are in progress, and results are expected in the near future. 93,102

Genetic Treatment of Systemic Diseases That Manifest With Cardiomyopathy Danon Disease
Danon disease is an X-linked multisystem disorder caused by dominant LoF variants in LAMP2. Danon disease emerges in affected men during childhood and typically causes death by late adolescence or early adulthood. 103 Because of X-inactivation, women can be affected but usually have milder disease. Cardiac manifestations include progressive and ultimately massive hypertrophy, markedly increased myocardial fibrosis, serious ventricular arrhythmias, and heart failure. 104 These phenotypes can be the presenting manifestation of Danon disease, 105 but clinical evaluations typically uncover skeletal myopathy and intellectual disability. Retinal, hepatic, and pulmonary involvement are also observed but less commonly. Alternative splicing of LAMP2 produces 3 transcripts (LAMP2A-C). LAMP2A and LAMP2B are ubiquitously expressed and are critical for chaperon-mediated (LAMP2A) 106 and macro (LAMP2B) autophagy 107 -a process in which lysosomes fuse with autophagosomes, 108 the vesicular packaging structure that contains cellular material targeted for autophagy. 109 Although the precise mechanisms by which pathogenic LAMP2 variants cause cardiac disease are unknown, deficits in autophagy may promote cell death, whereas premature glycogen accumulation may alter cardiac metabolism. 103,107,110 Because damaging LAMP2 variants cause loss of protein expression in hemizygous men, one treatment strategy would be viral-mediated delivery of LAMP2. Rocket Pharmaceuticals reported that AAV delivery of Lamp2 (RP-A501) improved cardiac, liver, and skeletal muscle phenotypes and survival in Lamp2-null mice and was safe in nonhuman primates. 111 Human clinical trials are planned for 2019.

Duchenne and Becker Muscular Dystrophy
Pathogenic variants in the DMD gene cause loss of dystrophin expression, resulting in Duchenne and Becker muscular dystrophies. 112,113 As DMD is encoded on chromosome X, damaging mutations are profoundly deleterious in hemizygous men, who present in childhood with progressive skeletal myopathy, which subsequently involves the heart and diaphragm, resulting in fulminant DCM and heart failure. Noncoding variants that disrupt physiological DMD expression cause X-linked disease in which cardiac phenotypes predominate. 114 Female carriers of pathogenic DMD variants can develop mild DCM later in life. 115 Pharmacological treatment goals in Duchenne muscular dystrophy aim to prevent micro-tears in skeletal muscles, stabilize membranes, and reduce inflammation. Although corticosteroids were once standard of care, more recent studies demonstrate that ACE inhibitors and mineralocorticoid antagonists 116 are more effective in maintaining skeletal muscle function 117 and delaying onset of cardiomyopathy. 116,118 Another emerging therapy aims to directly seal membrane tears due to the absence of dystrophin. The poloxamer P-188 NF is a chemical-based membrane sealant that inserts into the phospholipid bilayer of membranes improves muscle membrane integrity in preclinical models of muscular dystrophy, limiting the development of cardiomyopathy and associated fibrosis. 119,120 A phase II clinical trial is ongoing to evaluate the safety, tolerability, and efficacy in Duchenne muscular dystrophy patients. 121 The DMD gene is organized into 79 exons that span 2.5 Mb of genomic sequence that is spliced into a ≈4-kb transcript. 122 Because this large transcript size hinders the delivery of the entire gene sequence, minigene constructs have been developed and studied. AAV-mediated delivery of mini-DMD genes to mouse and canine muscular dystrophy models demonstrates some benefit but also evoked immunologic responses to the shorter peptide or AAV capsid 123 -an adverse effect that was minimized by repetitive injections of DNA vaccines, which is thought to induce tolerance. 124 An alternative approach to overcome the complexities associated with enhancing the expression of the large DMD gene is delivery of dual AAV constructs, each with half of the full sequence, along with appropriate splice donor and acceptor sequences to allow endogenous assembly of a large transcript. 125 Further therapeutic options emerged from detailed immunohistochemical studies of Duchenne patients and the Mdx mouse, which carries a frameshift variant in exon 20 and precludes translation of full-length dystrophin protein. Rare foci of muscle fibers were identified that express foreshortened but functional dystrophin because of spontaneous exon skipping. 126 These observations coupled with the recognition that many of the thousands of pathogenic DMD variants are clustered between exons 45 and 50 127 have propelled the therapeutic development of antisense oligonucleotides to promote exon skipping (Figure 3). AAV delivery of antisense oligonucleotides has successfully restored functional dystrophin in preclinical mouse and canine models. 128,129 In 2017, the Food and Drug Administration approved antisense oligonucleotidemediated exon skipping (Exondys 51, marketed as eteplirsen) for the estimated 13% of patients with pathogenic DMD variants that are amenable to exon 51 skipping. 130 More recently, using CRISPR/Cas9 and a single guide RNA, it was demonstrated that exon 51 could be deleted by targeting a region adjacent to the exon 51 splice acceptor site, allowing splicing from exon 49 to 52 and restoring the dystrophin open reading frame in over 70% of Ddm transcripts within 8 weeks. 131

Amyloidosis
Systemic amyloidosis defines a class of diseases caused by deposition of misfolded or misassembled proteins. Amyloidosis may present with systemic manifestations, but cardiac phenotypes are prominent in both immunoglobulin light-chain amyloid and transthyretin amyloidosis. The TTR gene encodes transthyretin-a homotetrameric carrier protein, which transports thyroid hormones in the plasma and cerebrospinal fluid. It is also involved in the transport of retinol (vitamin A) in the plasma by associating with retinol-binding protein. 132 Immunoglobulin light-chain amyloidosis is uncommon and occurs from plasma cell proliferative processes, including multiple myeloma. 133 TTR amyloidosis results from dissociation of stable transthyretin tetramers into intermediates that misassemble into misfolded monomers or oligomers that are deposited in the heart. 134 Both mutant transthyretin proteins containing 1 of ≈80 different dominant missense residues, or WT transthyretin protein, can cause cardiac amyloidosis. 135 TTR missense variants are more common among patients with African ancestry, younger age of disease onset, more systemic manifestation. These patients have a poorer prognosis. Cardiac manifestations occur in 80% of patients, typically after 60 years of age, and include concentric hypertrophy, heart failure, and conduction system disease, with survival limited to 2 to 3 years. 136 Patisiran-the first drug of its class-is an siRNA therapeutic that silences mutant TTR transcripts. It is also the first Food and Drug Administration-approved therapy for hereditary TTR amyloidosis. [137][138][139][140] Inotersen, another drug to recently gain Food and Drug Administration approval, is an antisense oligonucleotide that lowers hepatic levels of TTR. [141][142][143][144] Another pharmacological therapy has emerged from genetic analyses of familial TTR amyloidosis. Individuals with a heterozygous missense variant (TTR T119M) on one allele stabilized a prevalent pathogenic missense variant (TTR V30M) encoded on the other allele and suppressed disease development. 145 These observations led to screens that identified tafamidis and diflunisal-small molecules that stabilize transthyretin tetramers and inhibit their dissociation into amyloiogenic monomers. 146 A recent phase 3 trial ATTR-ACT (Transthyretin Amyloidosis Cardiomyopathy Clinical Trial) showed that tafamidis was associated with decreased all-cause and cardiovascular mortality, hospitalization, and functional decline in treated TTR cardiac amyloidosis patients, irrespective of genotype. 147

Remaining Challenges and Opportunities
The ultimate opportunity presented by discovering the genetic basis of cardiomyopathy is accurate prediction and disease prevention. The past decade has witnessed considerable advances to enable clinical progress, but many technical hurdles remain (Table). Somatic gene editing and transcript modification or correction is already technically feasible in cardiomyocytes, but adaptation of these strategies for in vivo human cardiac therapies will require considerable further scientific investigation, clinical data, and bioethical considerations.
A critical barrier to advancing manipulation of genes or transcripts relates to efficiency and off-target effects. Unlike self-renewing cells, in which the effects of somatic manipulation of genes in a few cells are amplified by transmission to daughter cells, cardiomyocytes undergo little replication. Hence, for therapeutic efficacy, a considerable proportion of cardiomyocytes will likely need to be targeted. Data are needed that define the proportions of treated cardiomyocytes that improve cardiac function, in addition to the development of highly effective vectors that can efficiently transduce many cardiomyocytes. These desired advances must be coupled with an equally important need to limit off-target events. Current approaches to detect off-target gene editing in the heart and other tissues are limited and require extensive sequencing of many cells-an impractical approach when vectors are systemically delivered. The delivery of exogenous genes, preprocessed RNAs for trans-splicing, silencing RNAs, and oligonucleotides to mediate exon skipping require the development of durable reagents or recurrent delivery mechanisms. Specifically, vectors are needed that have low immunogenicity, have the potential to accept large cargo, and enable systemic delivery with cardiac-specific activity.
Small molecules, while a traditional class of medicines that benefit from familiar strategies for delivery and ascertainment of efficacy, also have important limitations. The requirement for regular dosing is challenged by patient compliance, especially when drugs require multiple daily administrations to maintain effective blood levels. Small molecules can be nonspecific and despite extensive preclinical testing, many are found to cause unpredictable off-target effects that can lead to unacceptable clinical symptoms or outcomes. Small molecules can incite drug-drug interactions, especially when patients require multiple agents to manage cardiac symptoms or other medical conditions. Patient issues expand these multifactorial medicine complexities. For appropriate deployment of novel emerging genetic and small-molecule therapies, far more information is needed about the spectrum of genetic causes of cardiomyopathy. In most patients, these disorders emerge years after birth. Understanding the natural progression from the quiescent clinical state despite the presence of a pathogenic variant to the emergence of disease should inform the timing of interventions to maximize effectiveness and minimize risks. An appealing approach might be the delivery of therapeutics in advance of disease expression, but this raises many medical and ethical considerations. Longitudinal follow-up of some individuals harboring pathogenic variants demonstrates little or no manifestations of disease. Understanding the genetic, environmental, and lifestyle factors that influence disease penetrance could advance life-saving treatments in some and prevent un-needed risks and side effects in others. By contrast, delaying interventions until disease is fully established poses other complexities. Whether emerging genetic or small-molecule strategies can reverse established cardiomyopathy, in which altered contractility is associated with cardiomyocyte hypertrophy, atrophy, or demise, as well as myocardial fibrosis, remains unknown.
Advances that hold the potential for definitive therapies by somatic and germline gene editing require far greater consideration of ethical, legal, and societal issues. The engagement of stakeholders-patients, physicians, lay people, researchers, and policy makers-to educate and convene meaningful dialogue should promote responsible development and delivery of gene-editing therapies. The US National Academies of Science and Medicine, 148,149 the International Bioethics Committee of the United Nations Educational Scientific, 150 and Cultural Organization and the European Society of Human Genetics 151 have begun these discussions but to date have little inclusion of culturally diverse lay people and policy makers.
Further delineation of the genetic architecture of cardiomyopathies and elucidation of the molecular pathophysiology of pathogenic variants are expected to further propel the development of groundbreaking therapeutics. Undoubtedly, these discoveries will provide unparalleled insights into cardiac biology. In addition, these advances have enormous potential to attenuate disease, prevent adverse outcomes, and improve the lives of cardiomyopathy patients. With continuation of pioneering efforts in genetic manipulation and precision medicines for cardiomyopathies, deeper understandings are anticipated that will accelerate successful treatments and cures.

Sources of Funding
This work was supported, in part, with funds from the Stanley Sarnoff Foundation (G.G. Repetti), the Sir Henry Wellcome Fellowship, Wellcome Trust (206466/Z/17/Z to C.N. Toepfer), the National Institutes of Health (HL084553 and HL080494 to C.E. Seidman and J.G. Seidman), and the Howard Hughes Medical Institute (C.E. Seidman).

Disclosures
C.E. Seidman and J.G. Seidman are cofounders and own shares in Myokardia-a start-up company that is developing therapeutics that target the sarcomere. Myokardia had no role in this study and has not reviewed this article. The other authors report no conflicts.