Gene Therapy in Heart Failure

Abstract

With increasing knowledge of basic molecular mechanisms governing the development of heart failure (HF), the possibility of specifically targeting key pathological players is evolving. Technology allowing for efficient in vivo transduction of myocardial tissue with long-term expression of a transgene enables translation of basic mechanistic knowledge into potential gene therapy approaches. Gene therapy in HF is in its infancy clinically with the predominant amount of experience being from animal models. Nevertheless, this challenging and promising field is gaining momentum as recent preclinical studies in larger animals have been carried out and, importantly, there are 2 newly initiated phase I clinical trials for HF gene therapy. To put it simply, 2 parameters are needed for achieving success with HF gene therapy: (1) clearly identified detrimental/beneficial molecular targets; and (2) the means to manipulate these targets at a molecular level in a sufficient number of cardiac cells. However, several obstacles do exist on our way to efficient and safe gene transfer to human myocardium. Some of these obstacles are discussed in this review; however, it primarily focuses on the molecular target systems that have been subjected to intense investigation over the last decade in an attempt to make gene therapy for human HF a reality.

Heart failure (HF) (a cardiac condition that impairs the ability of the ventricle to fill with or eject blood) is the ultimate consequence of a vast number of cardiovascular diseases and constitutes 1 of the leading worldwide causes of morbidity and mortality. Pharmacological intervention with β-adrenergic receptor (β-AR) antagonists, inhibitors of angiotensin II and aldosterone, and diuretics are currently standard treatments for HF. Introduction of these pharmacological interventions have substantially increased survival and decreased morbidity in HF; however, these agents are far from ideal. As a consequence, there is an urgency to discover and develop novel and improved therapeutic strategies. Over the last 10 to 15 years, there has been a substantial increase in the understanding of cellular and molecular alterations that take place in HF, and this knowledge has resulted in the testing of hypotheses based on targeting molecular entities that appear to be involved in HF pathogenesis. Importantly, these scientific developments have provided a rationale for correcting human HF by directly intervening within the genetic/molecular foundation of the cardiomyocyte.

With the emergence and growth of translational cardiovascular research, the application of gene therapy for HF is no longer a futuristic “maybe.” However, to date, only 2 human clinical trials (both targeting SR Ca²⁺-ATPase [SERCA2a]) have been initiated with the aim of treating chronic human HF through gene delivery. These studies will be of critical importance to a few other genes/targets that are on the verge of human trials. Needless to say, this area of research is still in its infancy, and there are large numbers of signaling pathways that have potential pathophysiological significance, and these are currently at different stages of investigation. This review details the present status of HF gene therapy, emphasizing the most promising (and most explored) molecular target systems, as well as discussions regarding some prerequisites needed for successful human intervention.

Practical Considerations

Gene Therapy for HF: Searching for the “Golden Bullet”

In principle, gene therapy in HF must be aimed at correcting key molecular mechanisms in cardiac tissue that reduce/reverse the inevitable cardiac deterioration. This requires introduction of DNA/RNA that targets specific cardiomyocyte processes that alter HF outcomes. As such, prevention of lethal arrhythmias, acute cardiac decompensation, and prevention of end-stage pump failure must be achieved. This is a formidable task, and although it is unlikely that a simple “one target fits all” approach will be adequate, specific targets are being investigated and do appear to prevent or rescue HF in animal models. This single target approach is potentially possible because, although the etiologies off human HF are diverse, there are common biological and molecular signaling pathways within the failing cardiomyocyte that appear to exist. The ideal HF gene therapy strategy may depend on the specific cause of the condition (ie, ischemic, valvular, hypertensive, genetic, etc), perhaps requiring application of a combinatorial approach that targets different cardiac cells and signaling pathways. As an example, gene therapy for HF established as a consequence of ischemic heart disease may ideally involve: (1) targeting vascular cells by stabilizing coronary plaques and induction of neoangiogenesis; (2) providing means to prevent viable cardiomyocytes from dying (via inhibiting apoptosis and/or necrosis) and providing long-term inotropic support without adverse effects; (3) preventing or reducing myocardial remodeling by targeting fibroblasts; and (4) targeting myocardial electrophysiological abnormalities reducing the risk of arrhythmias.

Common to the above type of interventions is the need for a thorough basic understanding of the mechanisms governing specific abnormalities in order for effective therapeutic targeting. As an example of such targeting, the genetic intervention could be: (1) overexpression of a target molecule; (2) alteration of the target’s intracellular shuttling routes though introduction of decoy molecules; (3) loss of function approaches using dominant negative molecules or by introduction of RNA interference; (4) correcting deleterious gene mutations/deletions at the genome or primary mRNA level; or (5) installing genetically modified donor cells (stem cells and/or differentiated cells). In addition, appropriate interventional gene delivery techniques and vehicles/vectors need to be further developed to effectively achieve the above therapeutic approaches. Some of these topics are addressed in separate reviews within this series and are mentioned only briefly below. Our review herein focuses on means and targets to improve the function of the failing cardiomyocyte as a way to reverse myocardial dysfunction and remodeling.

Vehicles for Myocardial Gene Therapy in HF

Gene therapy in HF will certainly require efficient myocardial transduction and long-term transgene expression and only viral vectors appear to meet such a requirement. Even in settings where transient expression of a transgene may be sufficient (ie, induction of neoangiogenesis by secreted growth factors), results from larger controlled preclinical studies have been disappointing. This appears to be a consequence of low efficacy (see elsewhere¹ and references therein). A more detailed description of viral vectors and their application within cardiovascular disorders will be the topic of separate review. We discuss briefly below the most commonly used viral vehicles (adenoviruses [AdVs] and adeno-associated viruses [AAVs]) for HF gene therapy.

AdVs are easily manipulated, have a rather large transgene cloning capacity (≈7 to 8 kb), and can also be produced to high titers.² Importantly, in vivo experimental studies have demonstrated effective myocardial transduction with AdVs to alter global cardiac function.² The major disadvantage of AdVs is an inflammatory response generally seen on in vivo cardiac delivery, which culminates into a transient transgene expression and a secondary immune response if attempting to intervene again.³ Newer generations of AdVs devoid of more of the immunogenic viral epitopes (so-called “gutted” AdVs) may rectify this problem.

AAVs are generally considered more promising for gene therapy of chronic diseases such as HF because they readily infect cardiac tissue, produce stable and long-term transgene expression and are much less immunogenic.² Moreover, AAVs have not been shown to cause any known human disease.⁴ A major disadvantage of AAVs is their somewhat low packaging capability (≈4 to 5 kb). Furthermore, the presence of naturally occurring antibodies against some AAVs in the human population may limit their value.^4,5 Reports have also demonstrated unfavorable genomic integration sites and development of hepatocellular carcinoma in mice subjected to AAV gene therapy.⁶ A promising feature with AAVs is that some serotypes display tropism toward cardiac tissue,^7,8 and this may be exploitable to improve HF gene therapy strategies.

Gene Delivery Techniques

The ideal mode of HF gene delivery would be to use a vector that could be administered intravenously with efficient and specific uptake into cardiomyocytes. This could be an unreachable goal in humans because of the large blood volume and subsequent dilution issues; however, the cardiac muscle tropism of some AAV serotypes even in large animals^7,8 may prove that this approach is feasible. Overall, there is a concern about “scalability” of doses of vectors used in small animal models when attempting to translate into large animals and humans. Indeed, what has been used in mice and rats have been relatively large doses of viruses and it is not clear whether these will be sufficient to effectively transduce larger hearts or whether high titers potentially needed for human HF gene therapy are safe and even achievable. Importantly, the cardiac muscle tropism of some AAVs may lessen the amount of virus needed for effective human cardiac gene transfer.

Gene delivery strategies to the heart have been targeted interventions to the myocardium using either an intravascular approach through the coronary arteries or direct delivery to cardiac muscle. Intracoronary gene delivery is obviously clinically relevant and appealing; however, this approach is generally inefficient for myocardial gene transfer unless certain adjuvants and specific conditions are used (ie, increasing vascular bed permeability with agents such as histamine, vascular endothelial growth factor, or substance P and increasing perfusion pressure and allowing sufficient cell contact time for vehicle).^9,10 Importantly, some of these modifications may be challenging and problematic in larger animals and humans. However, there has been some success in larger animals where blocking venous return on coronary delivery of virus, or retrograde infusion of virus has been used.^11,12 In an attempt to further optimize intracoronary delivery, recirculation of virus by closed-loop systems is also being developed.¹³

Direct intramyocardial injection of viral vectors is also a technique that can support sufficient gene transfer including myocardial gene delivery in larger animals.¹⁴ The efficacy of this mode of cardiac gene delivery is naturally limited by accessible myocardial tissue and the areas injected. Moreover, some damage of tissue will occur along the needle track. However, this mode of vehicle delivery is conceptually appealing in human HF, especially during procedures where the myocardium is readily accessible (ie, cardiothoracic surgery).²

Promising HF Gene Therapy Targets

Targeting Proteins Involved in Cardiomyocyte Ca²⁺ Handling

The handling of Ca²⁺ during excitation–contraction (EC) coupling is a prominent feature of the cardiomyocyte. This important feature of the cardiomyocyte has long been a focus of molecular HF research because it is deranged in failing cells and importantly; this area will be targeted in the first human HF gene therapy trials. A detailed description of cardiac EC Ca²⁺ handling in normal as well as failing cardiomyocytes is beyond the scope of this review, but some points are important to address as they lay the groundwork for promising gene therapy targets. In brief, EC coupling begins with the initiation of a cardiomyocyte action potential and Ca²⁺ enters the cell through voltage-gated L-type Ca²⁺-channels. This initial influx of Ca²⁺ triggers the ryanodine receptor (RyR) to extrude Ca²⁺ from the sarcoplasmic reticulum (SR) into the cytosol.¹⁵ This Ca²⁺-induced Ca²⁺ release triggers cardiomyocyte contraction through Ca²⁺ binding to troponin C within the myofilaments of the sarcomere.¹⁵ Just as important is the removal of Ca²⁺ from the cytosol that initiates relaxation of the sarcomere. The cardiac SERCA2a and sarcolemmal Na⁺−Ca²⁺ exchanger (NCX) quantitatively are the major players in Ca²⁺ extrusion in larger animals.¹⁵ In addition to these nodal participants of EC coupling, molecules that regulate these proteins are critical for maintaining Ca²⁺-cycling stability.

Distorted handling of Ca²⁺ in the cardiomyocyte is a hallmark of HF. There is a decreased SR Ca²⁺ content and a prolonged Ca²⁺ transient, which is generally considered to be a consequence of increased NCX, a reduction of SERCA2a and a decreased phospholamban (PLN)/SERCA2a ratio, as well as an augmented open probability of the RyR (ie, “leakiness”).^16,17 Some divergences between these specific changes have been found depending the model of HF, as well as some species variations.¹⁸ In addition to these changes causing dysfunctional contractile performance, the decreased clearance of cytosolic Ca²⁺ may increase the risk of arrhythmias and also may precipitate pathological cardiac remodeling. ^17,19 Of importance, there is abundant data to suggest that targeting molecules involved in dysfunctional cardiomyocyte Ca²⁺ are attractive targets for HF gene therapy.

SR Ca²⁺-ATPase

SERCA2a is the cardiac isoform of this family of Ca²⁺-ATPases, and a loss of its activity and resultant decrease in SR Ca²⁺ uptake is a feature of the failing cardiomyocyte including in human HF.²⁰ SERCA2a activity in myocytes is tightly controlled by PLN, a small inhibitory peptide that regulates SERCA2a on a beat-to-beat basis. The inhibitory action of PLN is set by its phosphorylation status.²¹ When PLN is dephosphorylated it binds and inhibits SERCA2a. In contrast, when PLN is phosphorylated (primarily via protein kinase [PK]A but also calmodulin kinase II), its inhibitory activity on SERCA2a is alleviated. Conceptually, SERCA2a gene therapy in HF (replacing lost activity and tilting the PLN/SERCA2a equilibrium²²) is a rational venture as it would remove cytosolic Ca²⁺ faster in diastole and increase contractile reserves by increasing SR Ca²⁺ concentration. As an important proof-of-concept for human gene therapy, failing human cardiomyocytes showed restoration of impaired Ca²⁺ homeostasis and normalization of dysfunctional contractile responses after SERCA2a gene transfer.²³

The therapeutic potential of SERCA2a gene transfer to failing myocardial tissue has also been addressed in small animal models of HF. Intracoronary Adv-SERCA2a delivery to rats in HF resulting from transaortic constriction (TAC) led to improved systolic and diastolic function along with dramatically improved survival 28 days after gene delivery.^24,25 Importantly, these observations have been corroborated by findings in transgenic SERCA2a mice subjected to TAC.²⁶ The therapeutic potential of SERCA2a gene delivery has also been recently addressed in a large animal model of HF.²⁷ In this study, HF was induced by mitral valve impairment in pigs and AAV1-SERCA2a was administered by so-called antegrade epicardial coronary artery infusion on establishment of volume overload HF. Two months after SERCA2a gene delivery, clear improvements of left ventricular (LV) contractile performance and myocardial remodeling was observed. ²⁷ Consistent with the beneficial effects of increasing SERCA2a content to improve failing cardiomyocyte function is the opposite findings in heterozygote SERCA2a knockout (KO) mice subjected to TAC. In this experimental setting, an augmented hypertrophic response and a faster development of LV dilatation was observed.²⁸

Despite positive results in the above animal models, there are data from other studies that support a more tempered optimism for SERCA2a gene therapy. First, transgenic rats with cardiac SERCA2a overexpression were subjected to myocardial infarction (MI), and even though there were some beneficial effects on LV performance 1 month after MI, there were no differences between the groups when examined at 6 months.²⁹ Moreover, there was increased mortality seen in SERCA2a-treated rats within the first 24 hours after MI, which was apparently attributable to ventricular arrhythmias. ²⁹ In addition, gene transfer of SERCA2a to isolated failing canine cardiomyocytes, although rescuing diastolic dysfunction led to a loss of β-AR responsiveness and inotropic support.³⁰

Despite some of the above inconsistencies/concerns with SERCA2a gene transfer to cardiomyocytes, approval has been gained for the first human HF gene therapy trials. Two phase I clinical studies have been initiated where HF patients are to receive SERCA2a via myocardial gene delivery. In 1 study (CUPID—Calcium-Upregulation by Percutaneous Administration of Gene Therapy in Cardiac Disease—FDA approved, January 2007), intracoronary delivery of AAV1-SERCA2a (denominated MYDICAR; see http://www.celladon.net) will be given to HF patients. A report detailing the protocol design is now in press.^30a In brief, 1 part of the study is a dose-escalation study (4 doses per patient appropriately spaced in time) aimed at identifying safe and efficient dose of AAV1-SERCA2a delivered by antegrade epicardial coronary artery infusion to patients with ischemic or nonischemic dilated cardiomyopathy. Once the safe, efficient dose is identified, a smaller placebo-controlled randomized trial will be undertaken. Importantly, patients receiving AAV1-SERCA2a are required to also receive an implantable intra-cardiac defibrillator. Thus, the potential occurrence of ventricular arrhythmias discussed above will be addressed. The other phase I trial will deliver AAV6-SERCA2a to patients undergoing implantation of a left ventricular assist device with the primary goals of assessing safety and biological effects.

Phospholamban

Through its potent action on SERCA2a activity, PLN plays a nodal role in the regulation of SR Ca²⁺ homeostasis mediating slower cytosolic Ca²⁺ decay in cardiomyocytes, which translates into prolonged diastolic relaxation both in vitro and in vivo.^31,32 Increased levels of PLN in cardiomyocytes has been shown to be detrimental as higher levels of cardiac PLN overexpression in transgenic rabbits resulted in cardiomyopathy.³³ Transgenic mice overexpressing a mutant PLN “superinhibitor” of SERCA2a in cardiomyocytes also had deranged Ca²⁺ handling and depressed contractile performance leading to cardiomyopathy.³⁴ Consistent with these data are the findings that reduction of PLN levels (in KO mice) promote a dose-dependant hyperdynamic condition with increased cardiac performance and blunted contractile responses to β-AR stimulation (Figure 1 and Figure 2), the latter an apparent consequence of β-AR signaling being maximal.³⁵

Figure 1. Simplified β-AR and Ca²⁺ signaling in the normal, nonfailing heart under resting conditions.

(A) and after adrenergic stimulation (B). A, Sarcolemmal depolarization activates the L-type Ca²⁺ channel, and Ca²⁺ enters the cell, triggering the RyR to release Ca²⁺ from the SR (Ca²⁺-induced Ca²⁺ release). Elevated cytosolic Ca²⁺ binds troponin C and activates myofilaments resulting in contraction of the cardiomyocyte. During diastole, Ca²⁺ is removed from contractile proteins and the cytosol inducing relaxation of the sarcomere involving SERCA2a, NCX, and the mitochondrial Ca²⁺ uniporter. B, Adrenergic stimulation of β-ARs results in activation of AC, the generation of cAMP, and activation of PKA. The main targets of PKA include the L-type Ca²⁺-channel, the RYR, PLN, and troponin I, which, in concert, mediate positive inotropic and lusitropic responses. Adrenergic stimulation also induces GRK2 membrane translocation to initiate receptor desensitization.

Figure 2. Signaling in the failing heart.

(A), with emphasis on potential targets for gene therapy (B). A, Hallmarks of molecular signaling in the failing heart include defective β-AR signaling and distorted Ca²⁺ cycling. β₁-ARs are selectively downregulated, while β₁- and β₂-ARs are uncoupled from their effectors; both processes are mediated by GRKs, with GRK2 representing the predominant form in cardiac myocytes. GRK2 is significantly upregulated in the failing heart. Downstream, decreased SR Ca²⁺ content and a prolonged Ca²⁺ transient are generally attributed to increased NCX, decreased SERCA2a, decreased PLN/SERCA2a ratio, and leakiness of the RyR. B, Promising targets for HF gene therapy to normalize these changes include: (1) β₂-AR overexpression; (2) AC6 overexpression; (3) βARKct expression and GRK2 inhibition; (4) expression of a constitutively phosphorylated from of PLN; (5) overexpression of SERCA2a; (6) S100A1 overexpression; and (7) parvalbumin expression.

Being such a central player in the handling of Ca²⁺ and contractile function, the loss of PLN in various models of experimental HF has been studied. Cross-breeding PLN KO mice into 2 mouse models of genetic cardiomyopathy (MLP KO and CSQ-overexpressing mice) halted/rescued the detrimental phenotype.^36,37 However, in a study using a TAC model of mouse HF, smaller positive effects of PLN ablation was observed.³⁸ To also show that PLN ablation does not always lead to beneficial effects in HF, ischemic injury was exacerbated in isolated PLN KO hearts after ischemia/reperfusion.³⁹ Furthermore, somewhat neutral effects of PLN ablation in the development of cardiomyopathy and LV dysfunction have been found in some HF mice overexpressing tumor necrosis factor-α, Gαq, tropomodulin, or a mutant myosin-binding protein C.^40–43 Importantly, however, all of these studies do show the expected alterations in Ca²⁺ handling and contractility in isolated cardiomyocytes. Overall, these findings are interesting as they point out that genetic interventions promoting improved cardiomyocyte contractility and Ca²⁺ handling in HF cannot on a general basis predict a positive outcome in global cardiac function and remodeling. Moreover, these results point out that caution is needed in the interpretation of positive data in genetic mouse models of HF.

In attempts to translate the above findings of PLN lowering using gene therapy in experimental HF models, PLN inhibitory or dominant-negative molecules have been used. These studies have generally used mutant forms of PLN, such as a phosphomimetic mutation at serine-16 (S16E), the site of PKA action, which promotes SERCA2a activity.⁴⁴ Interestingly, AAV-mediated overexpression of PLN-S16E prevented cardiac deterioration in cardiomyopathic hamsters as well as in post-MI HF rats.^45,46 Lowering PLN levels by antisense technology has also been shown to correct contractility and Ca²⁺ handling in failing human cardiomyocytes.⁴⁷ Prevention of HF development in a larger animal has also been achieved by PLN-S16E gene therapy because sheep with cardiomyopathy demonstrated improved cardiac function compared to controls after intracoronary delivery of AdV-PLN-S16E.¹³

Even with a substantial amount of evidence showing that abolishment of PLN and/or removal of the inhibitory effects of PLN on SERCA2a can be beneficial in experimental HF, it is still unclear what the outcome would be if used in a human HF population. Of note, there are several documented cases of inherited human cardiomyopathies where the culprit has been found to be mutations and deletions in the PLN gene or its promoter and where the outcome is a loss of PLN inhibition on SERCA2a. Elegant studies have also shown that some of these human cardiomyopathies can be recapitulated in the mouse.^48,49 Moreover, a naturally occurring PLN “knockout” mutation in the human population is associated with severe cardiomyopathy.⁵⁰ This is a perfect example of the difficulties involved in extrapolating findings from murine models to the human population and shows that the enthusiasm for PLN-lowering in human HF should be tempered. However, PLN ablation genetically could be a very different biological event compared to a more acute intervention of lowering PLN levels in HF and probably more studies targeting PLN are warranted.

Protein Phosphatase 1 and Inhibitor Protein-1

The inhibitory action of PLN on SERCA2a is subject to tight secondary control mediated by protein phosphatase (PP)1 dephosphorylation. PP1 activity itself is also subject to rigorous control via the actions of the phosphatase inhibitors, inhibitor protein (I)-1 and I-2. β-AR–mediated PKA activation promotes I-1 to attenuate PP1 activity toward PLN, and this results in sustained PKA-mediated phosphorylation of PLN, which increases SERCA2a activity.⁵¹ Similar to murine gene manipulations targeting SERCA2a activity more directly, transgenic cardiac overexpression of PP1 or deletion of I-1 results in reduced cardiac performance, decreased β-AR-mediated Ca²⁺ signaling and contractility, and development of cardiomyopathy.⁵² Similar findings have been reported with the overexpression of the PP2 isoform.⁵³ I-1 levels and PP1 activity are reduced in human HF,⁵¹ and, thus, increasing I-1 levels in HF would constitute an appealing target. In an article by Pathak et al, this was specifically addressed in both mouse and rat models.⁵⁴ Transgenic mice overexpressing I-1 showed prevention of LV dysfunction following TAC, whereas beneficial cardiac effects were seen after intracoronary delivery of AdV-I-1 to rats with cardiac dysfunction after TAC.⁵⁴ These findings were corroborated in a recent study in which cardiac dysfunction in cardiomyopathic hamsters was halted by PP1 inhibition after AdV-mediated delivery of I-2.⁵⁵ Therefore, targeting PP1 and its inhibitory proteins to ultimately increase SERCA2a pump activity may constitute a promising gene therapy strategy in human HF, although there is clearly a need for further studies using different etiologies of HF and preferably larger experimental animal models.

Parvalbumin

Parvalbumin is an EF-hand Ca²⁺ sequestering protein that is exclusively expressed in fast-twitch skeletal muscle and neurons. It can bind 2 Ca²⁺ or Mg²⁺ ions per molecule, and the concentration and fluctuation rate of these ions determine the specific occupancy.⁵⁶ In contrast to overexpression of SERCA2a, parvalbumin introduction into cardiomyocytes offers an energy-independent removal of cytosolic available Ca²⁺ and, thus, in theory, constitutes a potentially appealing mode of correcting the prolonged diastolic Ca²⁺ decay generally seen in HF without further energy deprivation. As proof-of-principle, AdV-mediated expression of parvalbumin increased that rate of Ca²⁺ removal and improved the rate of relaxation in cardiomyocytes from hypothyroid rats, as well as in the heart after in vivo delivery.⁵⁷ Similar findings were seen after parvalbumin gene treatment of dysfunctional rat cardiomyocytes expressing a mutant α-tropomyosin (A63V), as well as in vivo using a cross-breeding strategy between α-tropomyosin-A63V transgenic mice and parvalbumin transgenic mice.⁵⁸ Selectivity of parvalbumin Ca²⁺ binding to the relaxation phase of the Ca²⁺ cycle is determined by the rate of Ca²⁺ influx and expression levels of parvalbumin.⁵⁹ Thus, there are some complex kinetic issues that could pose a potential problem in human HF gene therapy because differences in heart rate and/or expression levels of parvalbumin may cause Ca²⁺ sequestration during systole, which would compromise the force of contraction. This is the conclusion from a study of AdV-mediated parvalbumin to diastolic dysfunctional cardiomyocytes harvested from dogs subjected to TAC.³⁰ In this study, there was clear improvement of relaxation kinetics although at higher parvalbumin concentrations, sarcomere shortening was depressed.³⁰ Overall, although potentially promising, further investigations addressing the impact of long-term parvalbumin expression in relevant models of HF is warranted.

S100A1

S100A1 is a molecule that appears to play multiple and unique roles in cardiomyocyte Ca²⁺ cycling. S100A1 belongs to the S100 protein family (the largest EF-hand Ca²⁺-binding protein subfamily) and is predominantly and highly expressed in cardiomyocytes, where it localizes to the SR and the sarcomere and within the mitochondria.⁶⁰ Importantly, myocardial levels of S100A1 are decreased in HF.^60,61 S100A1 delivery to cardiomyocytes results in an increase of isometric contraction followed by an increase in the amount of Ca²⁺ pumped into the SR.^62,63 These actions of S100A1 are independent of the β-AR system because there are no changes in cAMP or PKA activity.⁶² However, β-AR stimulation in the presence of overexpressed S100A1 results in maximal contractile performance.^64,65

Interestingly, the regulatory targets of S100A1 at the level of the SR have been shown to be both RyR and SERCA2a. S100A1 stabilizes RyR in diastole (reducing the frequency of Ca²⁺ sparks) and augments Ca²⁺ release during systole.^61,66 Moreover, S100A increases SERCA2a activity during the relaxation phase.^60,66 Thus, in theory, S100A1 gene therapy for HF may offer a promising and novel mode of action that will not only promote beneficial cardiomyocyte EC Ca²⁺ handling but also potentially limit arrhythmias through stabilization of RyR and decrease in diastolic Ca²⁺ leak.

To date, animal models have illustrated the potential of S100A1 gene therapy for HF. First, S100A1 KO mice, although presenting with relatively normal cardiac function basally, display a substantial worsened LV function both after TAC and MI with significantly lower survival.^67,68 Importantly, the opposite result was seen in transgenic mice with cardiac-specific overexpression of S100A1 where LV function was stabilized after MI and inotropic reserve maintained. ⁶⁸ Moreover, AdV-mediated S100A1 gene transfer to failing rat cardiomyocytes resulted in restoration of disturbed Ca²⁺ handling through increased reuptake of SR Ca²⁺ during the relaxation phase and a lowering of the RyR-mediated Ca²⁺ leak, which also resulted in the reversal of fetal gene expression in this model.⁶⁹

A recent study has also shown chronic benefits of S100A1 gene delivery to the post-MI rat heart using an AAV vector.⁶¹ In this study, S100A1 myocardial gene transfer was performed using an AAV6 vector with a novel cardiomyocyte-specific promoter/enhancer (α-cardiac actin enhancer/elongation factor-1 promoter).⁶¹ Intracoronary delivery of AAV6-S100A1 was performed in rats 10 weeks after MI where significant LV dysfunction and HF was evident and 2 months after gene delivery, S100A1-treated HF rats presented with significantly enhanced cardiac function and a reversal of LV remodeling compared to control HF rats.⁶¹ This chronic rescue of HF was also evident at the single cardiomyocyte level, where restoration of intracellular Ca²⁺ transients and contractility was also seen.⁶¹ This study also had data supporting potential clinical importance because it showed additive beneficial effects combining S100A1 gene therapy with the β-AR blocker metoprolol.⁶⁰ Studies using large animal models as well as studies using human cardiomyocytes are definitely needed to continue to advance the idea that S100A1 addition could be a useful molecular therapeutic tool to combat human HF.

Targeting the β-AR System in HF

A hallmark of HF is alterations in cardiac β-AR signaling. It is now close to 3 decades since Bristow and colleagues demonstrated lower myocardial β-AR density and decreased responsiveness to β-agonists in failing human myocardium.⁷⁰ These observations triggered a vast amount of basic and clinical research investigating the molecular mechanisms responsible for these alterations. Several members of the β-adrenergic signaling pathway have been examined in respect to their role in maintenance of normal cardiac function, as well as their influence on HF development. Clinically, 2 pharmacological observations are clear: (1) use of β-AR antagonists in HF is generally beneficial⁷¹; and (2) sustained inotropic support in human HF through increased β-AR activation is detrimental.⁷² However, numerous basic investigations indicate that molecular manipulation of key players in the β-AR pathway that improve signaling and responsiveness can be beneficial in animal models of HF. This includes inhibition of receptor desensitization and increased adenylyl cyclase (AC) activity (see below). Naively viewed, these basic observations appear paradoxical considering the above-described clinical data, although the molecular complexity within the β-AR signaling pathway does not allow for a simplified “on is bad, off is good” interpretation.⁷³ Indeed, a more refined understanding of the balanced interplay of the different components of the β-AR signaling pathway may allow for novel molecular therapies that hold promise above and beyond what is currently being used clinically to improve cardiac function in human HF.

The complexity of the β-AR system in the heart begins at the level of the receptor. Both the β₁-AR and the β₂-AR are expressed in cardiomyocytes (see elsewhere⁷⁴ and references therein). In human myocardium, the β₁-AR accounts for 75% to 80% of total β-AR density.⁷⁵ In human HF, the β₁-AR is selectively downregulated (≈50%), and the remaining β₁- and β₂-ARs are partially uncoupled from their downstream signaling components including the heterotrimeric G protein Gs and AC.⁷⁵ This selective β₁-AR downregulation increases the importance and contribution of β₂-ARs for inotropy.⁷⁵ Importantly, recent data have shown that β₁- and β₂-ARs have distinct signaling and functional consequences, and, hence, they appear to regulate different aspects of catecholamine-dependent cardiac signaling and function.⁷³ Downstream of AC activation by β₁-AR stimulation, PKA is considered the key effector, whereas, recently, calmodulin kinase II has also been identified as an important cAMP-independent β₁-AR effector.⁷⁶ Of interest to this discussion, PKA has numerous intracellular targets, including key nodal regulators of EC Ca²⁺ homeostasis, discussed above such as PLN and RyR. In contrast to β₁-AR stimulation, β₂-ARs do not elicit the same myocyte contractile response, and the level of cAMP generation (at least globally) is negligible.⁷⁷

The interaction between activated β-ARs and their G protein is tightly regulated by kinases that dampen the receptor activity by phosphorylation of serine and threonine residues at the carboxyl terminus of the receptor.⁷⁸ This phenomenon is classically termed desensitization. Agonist-dependant desensitization (homologous desensitization) is mediated by a family of kinases known as the G protein– coupled receptor (GPCR) kinases (GRKs), whereas heterologous desensitization can occur with PKA and other kinases (ie, PKC).⁷⁸ The potential importance of GRK-mediated β-AR desensitization in HF is discussed below.

β₂-AR Overexpression

With the simple hypothesis that β-AR downregulation in the failing heart is maladaptive, studies in the 1990s used transgenic approaches to ask whether overexpression of β-ARs could alter the function and fate of the heart. The results of these investigations illustrate the complexity of cardiac β-AR signaling when pondering specific molecular manipulations to improve the function of the failing heart. Because dampening β-AR activity in theory may be considered an inborn protection mechanism for the failing heart (preventing arrhythmias and conserving energy), it was not surprising that transgenic mice with cardiac overexpression (30-fold) of the human β₁-AR caused severe cardiomyopathy.⁷⁹ In contrast, mice with cardiac overexpression of human β₂-ARs, at much higher levels than in the abovementioned β₁-AR mice, displayed enhanced systolic and diastolic performance without clear signs of deterioration.^80,81 β₂-AR–overexpressing mice also showed preserved contractility after MI.⁸²

However, there is cautionary evidence to warrant against β₂-ARs as a potential therapeutic molecule as transgenic mice with extremely high levels (>200-fold) of human β₂-ARs did develop fibrotic cardiomyopathy and HF after 40 weeks.⁸¹ Moreover, cardiac β₂-AR–overexpressing mice crossbred with HF mice overexpressing Gαq displayed augmented cardiac dysfunction and worsened hypertrophy.⁸¹ To increase the complexity of these findings and the “sorting out” of good versus bad β-ARs, a recent study in mice has shown that despite robust overexpression of the murine β₁-AR in the hearts of transgenic mice, no overt cardiomyopathic phenotype was observed.⁸³ The major difference between this mouse and the mouse described above⁷⁹ was the origin of the receptor (human versus mouse) and a possible significant signaling difference is the presence of constitutive receptor signaling activity (none for the mouse and present for the human).⁸³

The distinct difference between β₁- and β₂-AR overexpression in the hearts of transgenic mice and the detrimental versus beneficial phenotype has been attributed to the unique characteristic of β₂-ARs coupling to both Gs and Gi.⁷⁸ Importantly, studies have shown that whereas β₁-AR–mediated signaling leads to myocyte cell death including via apoptosis, β₂-AR stimulation leads to cell survival signaling combating apoptosis.^84,85

These results have led to testing whether raising β₂-AR levels in the failing heart may be protective or beneficial outside of the mouse and studies have used AdV-mediated β₂-AR gene delivery to larger animals. Supporting evidence was first found in rabbits because cardiac performance was improved after myocardial gene transfer of Adv-β₂-AR either via a global gene delivery method⁸⁶ or via intracoronary delivery.⁸⁷ AdV-β₂-AR cardiac gene transfer in a heterotopic rat heart transplant model improved basal, as well as zinterol-stimulated (a β₂-AR selective agonist), cardiac performance. ⁸⁸ Another gene therapy study with AdV-β₂-AR was its use in a heterotopic transplantation model in the rabbit, in which failing hearts were transplanted and β₂-AR overexpression accelerated reverse remodeling because of mechanical unloading following transplant and this included improved LV function in these failing hearts.⁸⁹

Although these studies provide intriguing data suggesting that β₂-AR enhancement via molecular upregulation may be beneficial, a note of caution is perhaps best served here because of conflicting reports (especially in the mouse). Furthermore, none of the above studies has examined long-term effects of myocardial β₂-AR overexpression in HF, including impact on survival.

Targeted Inhibition of GRK2

Studies over the last 2 decades have shown that limiting β-AR desensitization via GRK2 inhibition is a promising molecular strategy in HF.^73,78 The ubiquitously expressed GRK2 is the highest expressing GRK in the heart. GRK2 is a cytosolic protein, and upon GPCR stimulation, it is shuttled to the plasma membrane where it binds to dissociated and membrane-embedded βγ-subunits of heterotrimeric G proteins (G _βγ), allowing for phosphorylation of agonist-occupied receptors.⁷⁸ The in vivo effects of GRK2 on cardiac β-ARs was demonstrated in transgenic mice with cardiac-targeted GRK2 overexpression as these mice presented with blunted β-AR responsiveness both for AC activation and LV contractility. ⁹⁰ This is of potential clinical importance as GRK2 is upregulated in the failing human heart.⁹¹ The binding of G_βγ to GRK2 has been exploited experimentally as a peptide containing this region within the carboxyl terminus of GRK2, termed βARKct, has been used to inhibit GRK2-promoted β-AR desensitization, as well as other G_βγ processes.^73,78,92 Importantly, transgenic mice with cardiac-targeted βARKct expression display enhanced in vivo cardiac function⁹⁰ and reversal of the transgenic GRK2 phenotype.⁹³

The upregulation of myocardial GRK2 in human HF demonstrates that the decreased β-AR responsiveness typically found is probably attributable to augmented GRK2-mediated desensitization and receptor–G protein uncoupling. ^73,78 Importantly, numerous studies have also demonstrated similar upregulation of myocardial GRK2 in several experimental animal models of cardiac dysfunction. ^94–98 A decade ago, the dampening of β-AR signaling in HF caused by increased GRK2 was thought to be protective; however, data from our laboratory have shown that GRK2 activity in the failing heart appears to be pathological,^73,78 and its inhibition by the βARKct can lead to both the prevention and rescue of HF.^{93,95,96,99–104} This includes HF rescue in post-MI rabbits in which AdV-βARKct was administered via percutaneous intracoronary catheterization.¹⁰² As an important proof-of-concept for human HF, βARKct gene transfer to failing human cardiomyocytes significantly improved β-AR signaling and contractile dysfunction.¹⁰⁵ Of added interest, a study in βARKct transgenic mice with a murine model of HF showed an additive/synergistic effect on survival with concurrent administration of a β-AR antagonist,¹⁰¹ which demonstrates that the 2 strategies are not opposing. It is important to point out that because GRK2 acts to desensitize other GPCRs in the heart, the mechanism of the βARKct probably extends beyond its effects on β-AR signaling.

Importantly, neutral studies have also been reported using βARKct transgenic mice with some murine models of HF.^106,107 Moreover, a recent report shows that abolishment of GRK2 (using a newly developed conditional KO mouse for cardiomyocyte GRK2) precipitates cardiomyopathy on chronic catecholamine exposure.¹⁰⁸ Interestingly, we now have data using these GRK2 KO mice that show unequivocally that loss of GRK2 in cardiomyocytes preserves LV contractility, improves LV remodeling, and increases survival in post-MI HF (P.W.R., et al, unpublished data, 2008). The contrasting results between these 2 studies are probably a consequence of the model of LV dysfunction but also do show that caution is still warranted with GRK2 targeting. Overall, the data do support the use of the βARKct and GRK2 inhibition as a promising target for HF, and preclinical large animal studies are underway to potentially bring this molecule to clinical trials.

AC Type 6

There are 9 isoforms of AC expressed in mammals (AC1 through AC9¹⁰⁹), and AC5 and AC6 are the predominant cardiac isoforms.⁷⁷ Gene intervention promoting increased cardiac AC levels constitutes in theory an appealing venture because data indicate AC6 as rate-limiting in the β-AR/Gαs/ AC-mediated generation of cAMP. Perhaps more importantly is data demonstrating that overexpression of AC6 does not chronically and constitutively activate cAMP generation, as seen for β₁-ARs and Gαs-overexpressing transgenic mice.^106,110,111 In fact, cardiac AC6 overexpression leads to increased LV function and increased cAMP levels during β-AR stimulation with no basal changes.¹¹² This may constitute an important difference between using AC6 overexpression to improve β-AR–mediated cAMP production compared with enhancing signaling at the level of the receptor or G protein.¹¹³

The consequence of AC6 overexpression in HF has been tested experimentally in different animal models. Initial studies investigated the effect of cardiac AC6 overexpression in cardiomyopathic transgenic mice with myocardial overexpression of Gαq.^114,115 AC6 overexpression in the hearts of these mice led to improved LV function, reversal of dysfunctional β-AR signaling, and increased survival.^116,116 Importantly, these data have been corroborated in a more relevant gene therapy setting as Gq-overexpressing mice on intracoronary delivery of an AdV-AC6 vector had significantly improved cardiac function when studied 2 weeks after gene delivery.¹¹⁷

The beneficial cardiac effects of increased AC levels appear restricted to the AC6 isoform as overexpression of AC5 does not change outcomes with cardiac Gq overexpression. ¹¹⁸ In fact, the loss of AC5 (using KO mice) leads to improvements in cardiac function after TAC.¹¹⁹ Moreover, a recent report shows that AC5 KO mice have an increased lifespan and are protected against cardiac stress.¹²⁰ The apparent discrepancy between manipulating AC6 versus AC5 in the heart is, as yet, unresolved although it is interesting that cardiac AC5 overexpression (as opposed to cardiac AC6 overexpression) is associated with increased basal cAMP generation.¹²¹

One caveat to the above studies is that the Gq cardiomyopathic mouse model shows little resemblance to human HF. Thus, it is important to note that AC6 overexpression has been tested in other models of cardiovascular stress. For example, mice with cardiac overexpression of AC6 had a significantly lower mortality than control mice 1 week after induction of MI.¹²² Findings of significant increments in both isoproterenol and forskolin-stimulated cAMP production and LV contractility after intracoronary delivery of AdV-AC6 in normal, as well as pacing-induced, HF pigs has also been reported.^123,124 These studies show that the use of AC6 as a gene therapy strategy does hold some promise for use in HF and clinical trials have been planned. In theory, AC6 gene delivery could mediate inotropic support within the β-AR signaling pathway that, like inhibition of GRK2 with βARKct, would be temporally restricted without constitutive cAMP elevation.

Gene Targeting Monogenetic Causes of HF

Even though monogenetic causes of cardiomyopathy constitute a small percentage of the overall HF population, this diverse group of disorders potentially holds promise for more immediate gene therapy application. Being of monogenetic cause, they are (at least in theory) devoid of the complexity seen in most HF conditions. It is also hard to envision these disorders to ever be the focus of pharmaceutical drug development, and, thus, rectification of these disorders by gene therapy may be the only realistic treatment. The genetic basis for these mostly inherited diseases are widely heterogeneous and involve proteins constituting the sarcomere, the cytoskeleton, members of the Ca²⁺ handling family, and more.^125,126 For many of the cardiomyopathies caused by mutations in genes encoding sarcomeric proteins, it may be feasible to correct the altered function, and possibly prevent the development of HF, by simply overexpressing the wild-type protein. The rationale for this is the observation that some overexpressed, tagged exogenous sarcomeric proteins assemble into the sarcomere in an orderly, stoichiometric fashion. ¹²⁷ This presents the possibility of “removing” endogenously produced protein from the sarcomere by tipping the equilibrium. As of yet, no experimental studies have tested this theory; however, such experiments are feasible because mouse models that mimic known inherited human cardiomyopathies are available.

For cardiomyopathies precipitated by mutations in larger proteins, gene therapy is and will be much more technically challenging, if even possible. For example, Duchenne’s muscular dystrophy represents such a disease in which severe skeletal muscle weakness and cardiomyopathy are the major symptoms.¹²⁸ The mutant protein responsible is dystrophin and several different mutations or deletions in its large gene have been documented to give rise to a variety of different degrees of clinical manifestations.¹²⁹ In severe forms, the protein is significantly truncated or even absent. Importantly, it has been shown that introduction of a dystrophin minigene can substantially alleviate the clinical symptoms.¹³⁰ Genetic correction at the primary mRNA level by exogenously induced “exon skipping” has also been demonstrated to be efficient as this technology leads to the loss of exons harboring deleterious mutations during mRNA splicing resulting in production of a slightly truncated, but functional, protein.¹³¹ Using the same principle with antisense oligonucleotides, the feasibility of correcting Duchenne gene defects by exon-skipping was also recently demonstrated in human patients.¹³² Even though these results are promising, major challenges remain for treating both skeletal muscle as well as cardiomyocyte weaknesses in these patients.

Concluding Remarks

As more mechanisms involved in HF progression are uncovered, there will be an increasing number of molecules to potentially target in human HF. This review has focused on biological systems that have been explored the most over the last several years and those that have an ongoing emphasis. It is likely that with increasing basic understanding of cardiomyocyte processes that other targets will emerge that show equal or even superior promise than those identified to date. Obviously, all of the studies and data referred to in this review are primarily from experimental animal models that have a variable resemblance to actual settings seen in a human HF population. For the majority of human HF cases, etiologies are complex and a “one gene fits all” approach seems overly simplistic. Moreover, it is likely that when entering the human patient population, translation from experimental animal models will be difficult. Thus, it seems imperative that any candidate gene is subjected to a large amount of scrutiny including extensive studies in large animal models before attempting human trials. Further improvement in viral vectors and gene delivery methodology is also needed.

Despite these obstacles, gene therapy holds great promise in the aspect of improving survival and quality of life for HF patients. By specifically targeting molecular pathways known to be detrimental in HF, the stagnation currently seen for pharmacological therapies can be addressed. Importantly, a few of the targets discussed above are in preclinical stages with the goal of clinical trials in the near future. The initial safety and outcomes of the human trials ongoing with AAV-SERCA2a will be extremely valuable to fuel the journey of other targets and keep the momentum going for HF gene therapy.