Accumulating pre-clinical studies are providing an increasingly sound scientific basis for cautious human evaluation of additional gene therapy strategies. Furthermore, the anatomical compartmentalization of the heart and its accessibility by surgical and percutaneous approaches renders the myocardium a highly amenable target system for gene therapy. Improvement in our understanding of the molecular mechanisms of heart failure (HF), along with the development of novel and safer vectors for gene delivery, has led to the identification of novel targets that are difficult to manipulate pharmacologically but may be more amenable to gene therapy.

In this review, we highlight new strategies for the treatment of HF by gene transfer focusing on the vectors, targets, and delivery methods.

Nonviral vectors can be loosely grouped as plasmid DNA, liposome-DNA complexes (lipoplexes), and polymer-DNA complexes (polyplexes) (1). Oligonucleotides and their analogues, either alone or in complexes, are also an example of nonviral vector-mediated gene transfer. Although myocardial plasmid–mediated gene transfer is relatively inefficient (2), it has been the vector system on which several therapeutic angiogenesis trials have been based. In this setting, transient secretion of angiogenic factors by a modest number of gene-modified cells is sufficient for the desired phenotypic effect.

More recently, degradable polymer nanoparticles have been developed for gene delivery. Nanoparticle systems, polymers of hydroxy acids such as poly(lactic-co-glycolic) acid, which are U.S. Food and Drug Administration approved, have the capacity to release DNA material over prolonged time frames. Nanoparticles are well suited for gene delivery applications because they have a large nucleic acid cargo capacity and avoid the safety risks of standard adenoviral transfection (3). With the development of degradable polymer nanoparticles made of end-modified poly(β-amino ester)-amino ester that are easy to synthesize and can self-assemble with DNA to form stable nanoparticles that degrade via hydrolytic cleavage of backbone ester groups, there is now hope that these nanoparticles can be used in cardiac gene transfer.

The predominant use of viral vector systems in pre-clinical models of gene therapy is a reflection of the increased gene transfer efficiencies achievable with these systems. The most commonly used viral vector systems are covered in the following discussion, and in (Table 1), we list the 3 most commonly used vectors for cardiovascular gene transfer.

Recombinant human adenoviral vectors are the most commonly used vectors in experimental gene therapy models. The strengths of this vector system include the relative ease of production, the high functional vector titers achievable, and broad-target cell tropism, particularly within the cardiovascular system. All major cardiac cell types can be efficiently transduced by adenoviral vectors, both in vitro and in vivo. With regard to cardiomyocytes, efficient in vivo transduction has been demonstrated in gene therapy models of several mammalian species (4- 5). The adenovirus uses the Coxsackie adenovirus receptor to integrate into the cell. Physical approaches to enhance adenoviral gene transfer in the myocardium include methods that rely on catheter-based or surgically induced transcoronary pressure gradients to increase gene delivery to the myocardium (6) or the application of ultrasound energy to disperse the circulating vector as it traverses the myocardium (7).

Although adenoviral vectors are useful research tools, translation to clinical applications has been challenging for a number of reasons. Adenoviral vectors have a widely appreciated capacity for evoking intense immune and inflammatory reactions, which limits their expression. Additional constraints of these vectors include the promiscuity with which they infect all organs (especially the liver), even when delivered in a distant organ.

Several approaches have been explored to increase the biosafety of adenoviral vectors. These include vectors with increasingly reduced amounts of parental virus genes such that current generation “gutted” adenoviral vectors are devoid of almost all of these. This development results in decreased production of toxic viral proteins and may extend transgene expression over time.

Recombinant adeno-associated virus (rAAV) vectors are derived from the dependent parvovirus (8). This vector system has a number of clinically favorable attributes such as the lack of parental agent pathogenicity and vector-related cytotoxicity, minimal immunogenicity, and the capacity for stable long-term transgene expression through genomic integration and/or stable episome maintenance. Major limitations of rAAV vector systems include the production of high-titer vector stocks, a limited packaging capacity, and the existence of neutralizing antibodies in human populations.

AAVs have many characteristics that make them particularly well suited as vectors for cardiac gene transfer. First, AAV effectively transfects slowly or nondividing cells, making them ideal vectors for gene transfer to cardiomyocytes. Second, AAV is not known to cause human disease, suggesting that the risks to patients, health care workers, and the public at large are low. AAV vectors contain no viral genes, further increasing their safety, and decreased immune responses, thereby allowing them to persist and express transgenes in cells in the long term (>4 years in some trials [9]). A total of 12 different AAV serotypes are known, including AAV1 through AAV12, each with different tissue tropisms (10). AAV1 and AAV6 have been shown to transduce skeletal and cardiac muscle efficiently, whereas AAV9 has very high tropisms to the heart. Dimeric or self-complementary molecules (self-complementary adeno-associated virus [scAAV]) spontaneously re-anneal, alleviating the requirement for host-cell DNA synthesis. McCarty et al. (11) engineered such vectors and found that these scAAV resulted in rapid and higher levels of transgene expression than a conventional single-strand vector. These novel scAAV vectors represent a biochemical intermediate in rAAV transduction and should provide new insights into the biology of vector transduction. However, the cassette sizes must be half of a normal-size, single-strand vector (i.e., 2.3 kb), which further limits the number of genes that can be used with scAAV.

Limiting expression of the transgene to the heart is undoubtedly beneficial for clinical applications. One way to achieve expression exclusively in the heart is to use a cardiomyocyte specific promoter. However, although this is an attractive approach, it would be more desirable to achieve cardiomyocyte-specific expression by targeting the vector itself to the heart. Vector targeting will likely allow the use of lower vector doses compared with nontargeted AAV. Infusing lower vector doses are not only preferable from a production point of view but, more importantly, may be safer in patients. It was previously shown that chimeric viruses (i.e., virions composed of capsid proteins of >1 serotype) have distinct transduction profiles (12). Based on these results, viruses composed of hybrid capsid proteins (i.e., capsid proteins encoded by amino acid segments of several serotypes) should result in distinct tropism. To identify hybrid viruses that can selectively transduce cardiomyocytes in vivo, a combinatorial approach has been used whereby the generation of a library of diverse AAV variants, obtained by DNA shuffling, with an enrichment of cardiotropic AAV variants by directed evolution are combined (13).

Retroviral vectors based on the Moloney murine leukemia virus have been used widely in pre-clinical models. More recently, this vector was used in the successful French trial of gene therapy for X-linked severe combined immunodeficiency (14). The main limitations of retroviral vectors are an inability to transduce nondividing cells such as cardiac myocytes. As a result, many investigators have sought to overcome this block by inducing the in vivo proliferation in target cells or by using an ex vivo strategy (2).

Lentiviral vectors, based on the human immunodeficiency virus type 1 (15), transduce target cells by genomic integration. In contrast to retroviral vectors, these vectors are capable of transducing mitotically quiescent cells, a property that broadens the range of target cells, particularly within the cardiovascular system. Vector modifications addressing biosafety concerns associated with contemporary human immunodeficiency virus type 1–derived lentiviral vectors include the deletion of all accessory protein genes from the packaging system, separation of packaging elements into multiple plasmids, and the use of a chimeric 5′ long terminal repeat and a self-inactivating 3′ long terminal repeat in the vector plasmid (16). The experimental use of this relatively new vector system in pre-clinical models is rapidly expanding, and its potential application for human gene therapy has not yet been explored. The strengths of this system include the ability to confer long-term stable transgene expression. The major limitations relate to biosafety. Recently, premalignant clonal T-cell proliferation developed in 3 of 10 children from the French X-linked severe combined immunodeficiency trial that was directly attributable to dysregulation of a gene at the proviral integration site (17). The risk of insertional mutagenesis with integrating vector use had previously been considered minimal, a view now under revision.

The basic components of an expression cassette include promoter/enhancer elements, the gene(s) of interest, and an appropriate mRNA-stabilizing polyadenylation signal. Other frequently used cis-acting elements include internal ribosome entry site sequences to allow expression of ≥2 genes without the need for an additional promoter and introns and post-transcriptional regulatory elements to improve transgene expression.

Tissue-specific promoters can be used to restrict transgene expression to the desired target cell population and avoid unintended cells such as antigen-presenting cells. For example, cardiomyocyte-specific promoters such as alpha-myosin heavy chain have been used to restrict gene expression to the myocardium. Similarly, the smooth muscle–specific promoter SM22α has been demonstrated to restrict gene expression to cells of this type.

In a number of instances, it is desirable to have precise regulation of a therapeutic gene expression in vivo. Natural and synthetic enhancer-promoters can be used to drive gene transcription in a temporal, spatial, or environmental signal-inducible manner in response to heat shock, hypoxia, radiation, chemotherapy, or epigenetic agents. Hypoxia, intravascular shear stress, and left ventricular (LV) strain have all been used in models of this type of regulation. In the setting of HF, a regulatable system would be important if the gene of interest needs to be turned on or off for a short period of time. This would include genes driving angiogenesis or stem cell recruitment or expansion.

Meganucleases and zinc finger proteins can be engineered to induce double-strand breaks at specific DNA sequences (18- 19). These breaks are repaired by homologous recombination or by nonhomologous end-joining, which results in insertions or deletions of a few base pairs. They can then be used to restore the normal reading frame of a gene with a specific mutation. Similarly, engineered zinc finger protein transcription factors can specifically activate or repress virtually any gene. Meganucleases and zinc finger proteins have been used to target dystrophin mutations and vascular endothelial growth factor expression by gene transfer (20). With the ever-expanding genetic mutations in dilated and hypertrophic cardiomyopathies (21), these nucleases can be used in the future to repair DNA in the affected organs by gene therapy.

In addition to the major advances in engineering vehicles for myocardial gene therapy, novel delivery methods have been developed to specifically target the heart. Regardless of the delivery methods, chemical approaches that include the use of vasodilatory and permeabilizing agents have been used to facilitate transfer of vector from the vascular lumen to the myocardium (22). In fact, several agents that increase the permeability of the vascular bed have been used in pre-clinical trials including nitroglycerin, nitroprusside, serotonin, bradykinin, histamine, substance P, and vascular endothelial growth factor (23). Clinically and in the setting of HF, these agents must be used with caution so as not to decrease systemic blood pressure. The multiple methods of gene delivery are illustrated in (Figures 1, 2).

Antegrade injection in the coronary circulation is the most straightforward way to deliver vectors to the myocardium. It has the potential to homogeneously deliver the vector to the whole myocardium. There are multiple techniques of antegrade coronary injection.

This method, depicted in (Figure 1)A, involves inflating a balloon to block the coronary flow while injecting downstream of the balloon. This method allows flow of the viral vector to occur without dilution; however, blocking antegrade flow may not be well tolerated, even for a short period of time.

This simple yet effective technique, depicted in (Figure 1)B, works best with AAV vectors. There is no obstruction of flow, and it can be an optimal method for patients with HF who may not tolerate coronary artery blockade. Even though this method of antegrade injection does not infect all cardiomyocytes (∼60% of myocytes), it follows the normal pattern of flow in the coronary arteries and has been shown to improve ventricular function in a large animal model of HF (24). Of note, this is the method that was used clinically in patients with severe HF receiving AAV1 sarcoplasmic reticulum Ca²⁺ ATPase (SERCA2a).

The V-Focus (Osprey Medical Inc., St. Paul, Minnesota) delivery system allows a minimally invasive, percutaneous procedure to establish isolated regional perfusion of the myocardium (Figure 1C). It enables a closed circuit to be percutaneously established between the coronary arteries and the coronary sinus. Through percutaneous catheters, it selectively isolates and circulates the gene transfer agent through the coronary circulation. Adequate oxygenation of the myocardium is maintained in part by circulating the perfusate through an extracorporeal membrane oxygenation system. Once a closed circuit is established, the vector is directly injected into the circuit and circulates through the myocardium for approximately 10 min. This technique has allowed widespread transduction of SERCA2a in an ovine model of HF with a substantial increase in contractility, both with adenoviral and AAV gene transfer (25- 26).

Using the coronary vein, which is disease free in most cases, for the route to deliver the therapeutic material is attractive, especially in the clinical setting in which the patient has a diseased coronary artery (Figure 2A) (27- 29). The increase in efficacy has been achieved by pressure-regulated retroperfusion using a specialized device in this approach. Retrograde infusion of vector via the coronary veins constitutes a novel catheter-based technique for myocardial gene delivery, and in studies pioneered by Boekstegers et al. (27), it was shown to be superior to antegrade coronary injections. However, other studies showed that antegrade coronary injection was superior to retrograde injection (30). However, this method can be somewhat problematic in patients who may not tolerate coronary occlusion.

A composite catheter system (TransAccess, Medtronic Inc., Minneapolis, Minnesota) incorporates a phased-array ultrasound tip for guidance and a sheathed extendable nitinol needle for transvascular myocardial access that is advanced from the coronary sinus retrograde (Figure 2B). A microinfusion catheter is advanced through the needle, deep into remote myocardium, and the autologous cell-hydrogel suspension is injected into normal heart. This allows the delivery of vectors directly into specific areas of the myocardium accessible through the coronary sinus (31).

Direct injection of the vector into the myocardium can be achieved surgically or percutaneously. It overcomes numerous potential drawbacks that can be encountered with the intravascular route including the first-pass effect of the liver and spleen, the effect of neutralizing antibodies, the T-cell response, and the impermeability of the endothelial barrier. Surgical gene transfer can be performed via a subxiphisternal or transthoracic approach and can be an attractive approach when the myocardium is easily accessible as during cardiothoracic surgeries (Figure 3A). Catheter-based needle endomyocardial injection can be performed with several types of catheters and under different guidance modalities including fluoroscopy, electromagnetism, and 3-dimensional mapping systems (Figure 3B). Both surgical and percutaneous approaches have limited vector delivery secondary to the restricted area of injection and the leakage at the site of the myocardial injection.

The pericardial sac is a closed space in close proximity to the myocardium and is accessible to both surgical and percutaneous delivery (Figure 3C). Vectors in this space preferentially transduce the pericardial cells with minimal myocardial expression. A percutaneous pericardial approach allows safe and effective pericardial access appropriate for the delivery of therapeutic agents.

The optimal method of delivery depends largely on the proportion of target tissue to be genetically modified. Focal transduction might be desired to salvage an ischemic area, whereas diffuse gene expression is more likely to reverse global myocardial dysfunction.

Surgical delivery in animal models usually involves an open chest procedure in which the heart is arrested in hypothermia or the aorta and pulmonary arteries are cross-clamped in rodent models (6,32- 33) or the animals are placed on bypass in large animal models. On bypass, the viral vector can be delivered in either an antegrade or retrograde fashion (34). More recently, Bridges et al. (35- 36) developed a complete surgical isolation of the heart in situ with retrograde administration of vectors, which results in very high myocyte transduction efficiency. However, it is not clear whether patients with advanced HF could tolerate such a procedure for gene transfer.

The past 20 years witnessed significant evolution in our understanding of the pathophysiology of HF in its molecular and cellular dimensions, which broadened the scope of interventions available for gene therapy. We discuss in this section some of the most important systems targeted to restore the function of failing cardiomyocytes.

For targets to be validated, it is important that they rescue function in animal models when HF has been already established, that the rescue is not associated with arrhythmogenesis, and that a gene-dose effect is established (i.e., with increasing expression of the gene of interest, there is a concomitant improvement in function).

In (Figure 4), the various targets in excitation contraction coupling are presented.

The β-adrenergic system is affected by multiple alterations including β-adrenergic receptors (β-ARs) down-regulation, up-regulation of β-AR kinase, and increased inhibitory G protein function. These alterations lead to desensitization of the β-ARs and decreased signaling through their pathway. Several gene-based experiments tested the hypothesis that genetic manipulation of the myocardial β-AR system can enhance cardiac function.

Overexpression of β-AR was initially tested as a simple way to overcome β-AR down-regulation. Transgenic mice overexpressing human β₁-ARs had severe cardiomyopathy (37). This finding reinforces the hypothesis that β-AR down-regulation is a protective mechanism in the failing heart. In contrast, mice with cardiac overexpression of β₂-AR demonstrated increased basal myocardial AC activity with increased LV function (38). Both direct and intracoronary myocardial delivery of adenovirus containing the human β₂-AR transgene resulted in enhanced cardiac performance in rabbits (39- 40).

The interaction between activated β-ARs and G proteins is regulated by kinases that modulate the receptor activity by phosphorylation of its carboxyl terminus. Agonist-dependent desensitization is mediated by a family of GRKs that phosphorylate the agonist-occupied receptors, resulting in functional uncoupling. GRK2 is the most expressed GRK in the heart. It has been implicated in the pathogenesis of dysfunctional cardiac β-AR signaling, accounting for a deleterious activity in the failing heart (41). Studies in mice in which HF was induced by a myocardial infarction showed that selective GRK2 ablation 10 days post-infarction resulted in increased survival, halted ventricular remodeling, and enhanced cardiac contractile performance (42). A peptide termed βARKct capable of inhibiting GRK2-mediated β-AR desensitization has been evaluated in vivo in animals. Using intracoronary adenovirus-mediated βARKct transgene delivery to rabbits 3 weeks after induced myocardial infarction demonstrated a marked reversal of ventricular dysfunction (43).

Although detrimental outcomes were demonstrated with multiple elements of the β-adrenergic system used to improve the expression of cyclic adenosine monophosphate, activation of AC type VI seems to have a unique favorable profile. Overexpression of AC type VI in transgenic mice resulted in improved cardiac function in response to adrenergic stimulation along with increased cyclic adenosine monophosphate production in isolated cardiac myocytes. Importantly, AC type VI had a neutral effect on basal heart function and was not associated with any structural heart abnormalities (44). In a pacing model of HF in pigs, intracoronary delivery of adenovirus encoding AC type VI resulted in improved LV function and remodeling, associated with increased cyclic adenosine monophosphate–generating capacity (23). The favorable effects of AC type VI in pre-clinical studies are encouraging, and this approach is currently under investigation for initiation of clinical trials in patients with HF (45).

Ca²⁺ plays a crucial role in contraction and relaxation phases of the cardiac cycle (Figure 1). Ryanodine receptors are proteins forming a link between the T tubules in the cardiomyocytes and the sarcoplasmic reticulum (SR). In a process known as Ca²⁺-induced Ca²⁺ release, depolarization activates voltage-operated l-type Ca²⁺ channels of the T tubule to allow Ca²⁺ entry into the cardiomyocyte. Ca²⁺ reaches the ryanodine receptor protein, modifying its molecular configuration, which in turn opens the Ca²⁺ release channel of the SR, releasing Ca²⁺ into the cytosol. This process greatly increases the concentration of Ca²⁺ in the cytosol, allowing it to interact with troponin C, triggering the contraction process. Relaxation occurs when Ca²⁺ detaches from troponin C and is either taken up by the SR via the action of SERCA2a or extruded from the cardiac cell by the sarcolemnal Na⁺/Ca²⁺ exchanger. Phospholamban (PLN) is a protein implicated in the regulation of SERCA2a activity. In its unphosphorylated form, PLN can inhibit SERCA2a. In contrast, phosphorylation of PLN relieves its inhibitory effect and results in increased SERCA2a activity with improved Ca²⁺ handling. Protein phosphatase 1 (PP1) is a type of serine/threonine phosphatase that has a major role in dephosphorylation of PLN in the heart (46- 48). In particular, stimulation of the β-adrenergic axis induces phosphorylation of a phosphatase protein inhibitor 1 (PPI) resulting in PP1 inhibition and enhancement of myocardial contractile function (46- 48). HF is characterized by multiple defects in Ca²⁺-handling proteins with impaired SR Ca²⁺ loading and release. Reversal of those defects by gene therapy techniques has shown very promising results. We review the main aspects of those novel therapies in this section.

More than 20 years ago, Gwathmey et al. (49) first reported that calcium cycling is abnormal in human HF and was found to be partially due to decreased SERCA2a activity regardless of the etiology of the HF (50- 52). Improvement in cardiac contractility after gene transfer of SERCA2a has been demonstrated in a large number of experimental models of HF (54- 55). More importantly, long-term overexpression of SERCA2a by intracoronary delivery of AAV carrying SERCA2a has been associated with preserved systolic function and improved ventricular remodeling in a swine volume-overload model of HF (24). Beyond their effects on enhancing contractility, SERCA2a gene transfer has been shown to restore the energetics state of the heart both in terms of energy supply and utilization to decrease ventricular arrhythmias and to enhance coronary flow through activation of endothelial nitric oxide synthase in endothelial cells (56- 59).

Another approach to improve Ca²⁺ handling involves inhibition of PLN. Decreasing PLN in human cardiac myocytes showed an improvement in contraction and relaxation velocities similar to the benefit seen with gene transfer of SERCA2a (60). Silencing of PLN expression in a sheep HF model resulted in improved SERCA activity along with improved systolic and diastolic LV function (26). In addition to these conventional gene therapy strategies, RNA interference therapy was used for the first time in a model of cardiac disease, specifically in rats with HF, in an attempt to suppress PLN expression. An rAAV-RNA interference vector generated stable cardiac production of a regulatory RNA sequence, which in turn suppressed PLN expression. SERCA2a protein was subsequently increased accompanied by restoration of systolic and diastolic cardiac function (61).

HF is associated with elevated PP1 activity in humans resulting in dephosphorylation of PLN. Overexpression of PP1 or ablation of PPI in murine hearts has been associated with decreased β-AR–mediated contractile responses, depressed cardiac function, and premature death consistent with HF (46- 48). Expression of a constitutively active PPI in transgenic mice led to PP1 inhibition with increased phosphorylation of PLN and improved cardiac contractility. A recent study on transgenic mice expressing active PPI confirmed the relationship between phosphorylation of PLN and SERCA2a activity. PPI expression ameliorated ischemia/reperfusion-induced injury by reducing the infarct size and improving contractile recovery in addition to decreasing biomarkers of apoptosis and endoplasmic reticulum stress response (46- 48).

S100 is a multigenic family of Ca²⁺-modulated proteins implicated in intracellular and extracellular regulatory activities. S100A1 is the most abundant S100 protein isoform in the heart. It promotes cardiac contractile and relaxation function through enhancing the activity of both ryanodine receptors and SERCA2a (62). In a rat model of HF, AAV6-mediated long-term expression of S100A1, resulted in a sustained in vivo reversal of LV dysfunction and remodeling (63- 64).

Apoptosis is a process of programmed cell death that is involved in normal organ development. In animal models of acute and subacute ischemia/reperfusion, overexpression of the antiapoptotic protein Bcl-2 or akt reduces the rate of cardiomyocyte apoptosis and improves heart function (49). Despite the early encouraging results in animal studies, antiapoptotic strategies face multiple challenges before being considered for human trials. There remain some uncertainties about the reliability of the techniques currently used to estimate the rate of apoptosis (53). In addition, the loss of the normal cellular regulation role provided by apoptosis can generate a serious risk of malignant cellular proliferation and autoimmune injuries. Finally, although ischemia/reperfusion injury seems to be amenable to intervention, it is less clear whether other forms of cardiac injury such as hypertrophy and HF can benefit from antiapoptotic strategies.

It has been well established that damaged and inflamed tissues send out signals to attract stem cells to the injured tissue, and many of these signals have been identified, including stromal-derived factor (SDF)-1. There has been growing evidence that SDF-1:CXCR4 (its receptor) axis plays an important role in regulating myocardial repair after ischemic injury. SDF-1 has been shown to be increased after myocardial ischemia, and several studies have shown that enhancing SDF-1 levels around the infarct improves myocardial remodeling after infarction. Currently, the strategy of delivering SDF-1 naked DNA by percutaneous measures to the peri-infarct area is being used clinically (65).

Once a target is validated in cells, rodents, and large animal models; the appropriate vector has been identified; and a route of delivery is chosen, a number of challenges remain before starting clinical trials. These include evaluation of the immune response, clinical parameters to be examined, and evaluating the risk of arrhythmias.

One of the challenges with viral gene transfer is the pre-existence of neutralizing antibodies. In different studies, it has been shown that about a significant proportion of adults are seropositive for AAV2 (as much as ∼80%), AAV1 (∼50%), AAV5 (∼40%), and AAV6 (∼30%) (66- 67). The presence of pre-existing antibodies can have considerable implications for cardiovascular gene therapy because it has been shown in several studies that these AAV-specific antibodies are neutralizing (66- 67) and can severely compromise the utility of this approach. These results highlight the necessity to determine neutralizing antibody titers against the specific vector used in the patient population.

Another issue that can occur is that cells may transiently express AAV capsid protein on their cell surface. T-cell response could occur in any organ, but the greatest concern is for the liver or the heart (site of injection). This response is dose dependent, and in clinical trials to treat hemophilia or lipoprotein lipase deficiency, higher doses (>10¹³ viral genomes) were associated with activation of capsid-specific T cells and elimination of the transgene (68- 69). In clinical trials, to evaluate the potential development of a T-cell response, it is important to use an ELISPOT assay to detect anti-AAV capsid T-cell responses (interferon-γ release when a patient's peripheral blood monocytes are exposed to capsid peptide).

It is critical that the assessment of the efficacy of gene therapy trials be robust enough to detect biological signals. In pre-clinical trials and large animal studies, investigators have relied mainly on hemodynamic measurements coupled with noninvasive techniques such as echocardiography and magnetic resonance imaging (MRI). MRI, which provides the most detailed structural measurements of the heart, cannot be routinely used in advanced HF patients because a majority of them would have implantable cardiac defibrillators or biventricular pacemakers. Clinically, invasive monitoring is obviously difficult to perform; however, there are a number of clinical measures that can be followed.

Arrhythmias have been clearly shown to be a consequence of pharmacological inotropic therapy. In the context of gene therapy, there may be a number of reasons for arrhythmias to occur. The overexpression or knock down of specific genes such as channels and ionic transporters may induce changes in electrical remodeling that can predispose the cells to arrhythmias. In addition, inhomogeneous expression within the myocardium can induce areas of heterogeneously contracting cardiomyocytes that can also lead to arrhythmias. It is therefore essential that in pre-clinical studies, detection of arrhythmias be thoroughly studied after gene transfer using Holter monitors or implantable continuous monitoring. Clinically, it would be important to protect the patient receiving the gene therapy product with an implantable cardiac defibrillator.

Despite the initial hope that gene therapy had at its inception, the failures of the clinical trials coupled with the serious adverse events brought the field close to collapse 10 years ago. The well-publicized complications included the death a patient with a mild form of the monogenic disorder ornithine transcarbamylase deficiency, which was caused by multiorgan failure after the injection of a recombinant adenovirus (70- 72), and the development of new T-cell lymphomas in 3 of the 10 children with severe combined immunodeficiency after treatment with retroviruses (14,73- 75). However, the field of gene therapy learned valuable lessons from these failures and has rebounded recently with rigorous regulatory barriers along with a new wave of novel vectors. There is now a growing number of gene therapy trials for various diseases, most notably inherited blindness (whereby gene transfer by AAV vectors partially restored vision in a pediatric patient with Leber's congenital amaurosis, a major cause of congenital blindness [76]), cancer, infectious diseases, monogenic diseases, and cardiovascular diseases.

Specifically, in HF, there are currently a number of trials ongoing or in the planning stages targeting various pathways for rescuing the failing myocardium (Table 2). The targets that have been taken forward for clinical trials include SERCA2a, AC type 6, and SDF-1.

The first clinical trial of gene therapy in patients with HF was launched in the United States in 2007 (77- 78). CUPID (Calcium Up-Regulation by Percutaneous Administration of Gene Therapy in Cardiac Disease) is a multicenter trial designed to evaluate the safety profile and the biological effects of gene transfer of the SERCA2a cDNA by delivering a recombinant AAV1 (AAV1.SERCA2a) in patients with advanced HF. Participants in this trial were administered a single intracoronary infusion of AAV1.SERCA2a using an open-label approach (77- 78). Cohorts 1 to 4 of 3 patients each received sequentially a single increasing dose of AAV1.SERCA2a. The 12-month follow-up of these patients showed an acceptable safety profile (77- 78). Improvement was detected in several patients, reflected by symptomatic (n = 5), functional (n = 4), biomarker (n = 2), and LV function/remodeling (n = 6) parameters. Although this was a phase 1 study involving a small number of patients, early results found that AAV1.SERCA2a treatment conferred a quantitative biological benefit. In the phase 2 trial, 39 patients with advanced HF were randomized to receive intracoronary AAV1-mediated SERCA2a gene delivery (in 1 of 3 doses: low dose, 6 × 10¹¹ DNAse-resistant particles; middle dose, 3 × 10¹² DNAse-resistant particles; and high dose, 1 × 10¹³ DNAse-resistant particles) versus placebo. Patient's symptoms (New York Heart Association functional class, Minnesota Living With Heart Failure Questionnaire, functional status [6-min walk test, and Vo₂max]), N-terminal pro–B-type natriuretic peptide levels, and echocardiographic measures were evaluated over 6 months. Treatment success was determined by examining concordant trends in these end points for group- and patient-based comparisons as well as clinical outcomes. The AAV1.SERCA2a high-dose group met the prespecified criteria for success at the group and individual patient levels. AAV1.SERCA2a-treated patients, versus placebo, demonstrated improvement or stabilization in New York Heart Association functional class, Minnesota Living With Heart Failure Questionnaire, 6-minute walk test, Vo₂max, N-terminal pro–B-type natriuretic peptide levels, and LV end-systolic volumes. Significant increases in time to adjudicated cardiovascular events and a decreased frequency of cardiovascular events per patient were observed in all patients receiving AAV1.SERCA2a. No increases in adverse events, disease-related events, laboratory abnormalities, or arrhythmias were observed in AAV1.SERCA2a-treated patients compared with those receiving placebo.

Two other clinical trials targeting SERCA2a are currently enrolling patients. The first trial is in patients with advanced HF having received LV assist devices at least 1 month before treatment and who will receive either AAV6.SERCA2a or saline solution. This trial is being conducted in the United Kingdom. A second phase 2 single-center, double-blind, randomized, placebo-controlled, parallel study will be held at the Pitié-Salpêtrière Hospital Institute of Cardiology, Paris, France, with the primary objective to investigate the impact of AAV6-CMV-SERCA2a on cardiac remodeling parameters in patients with severe HF.

In a separate clinical study, AAV5 encoding human AC type 6 is being delivered by intracoronary injection to patients with congestive HF. Intracoronary delivery of Ad5.hAC6 or phosphate-buffered saline in 3:1 randomization with dose escalation, The patients will be randomized in a dose-dependent fashion starting at 3.2 × 10⁹ viral particles to 3.2 × 10¹² viral particles in 6 dose groups in a 3:1 randomization (with phosphate-buffered saline being used for control). The trial is currently enrolling patients.

An additional trial is examining the effects of injecting SDF-1 directly into the myocardium of patients with ischemic heart disease. An open-label dose escalation study to evaluate the safety of a single escalating dose of SDF-1 administered by endomyocardial injection to cohorts of adults with ischemic HF is currently enrolling patients. SDF-1 naked DNA will be injected directly into the myocardium at multiple sites via a percutaneous, LV approach.

An increase in our knowledge of molecular mechanisms of HF along with improved gene therapy technology has led to substantial efforts in pre-clinical testing of a number of targets and more recently in the successful completion of a phase 2 trial in gene therapy for HF. Now that the safety of AAV vectors has been established for the treatment of HF along with the efficacy of SERCA2a in the treatment of HF, the field is now open for testing novel targets that are pharmacologically difficult to modulate with more advanced AAV-based vector systems.