Gene and cell-based therapies for heart disease

Gene and cell-based therapies for heart disease LUIS G. MELO,*,,1 ALOK S. PACHORI, DELING KONG, MASSIMILIANO GNECCHI, KAI WANG,* RICHARD E. PRATT, AND VICTOR J. DZAU *Department of Physiology, Queen s University, Kingston, Ontario K7L 3N6, Canada; and Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts USA ABSTRACT Heart disease remains the prevalent cause of premature death and accounts for a significant proportion of all hospital admissions. Recent developments in understanding the molecular mechanisms of myocardial disease have led to the identification of new therapeutic targets, and the availability of vectors with enhanced myocardial tropism offers the opportunity for the design of gene therapies for both protection and rescue of the myocardium. Genetic therapies have been devised to treat complex diseases such as myocardial ischemia, heart failure, and inherited myopathies in various animal models. Some of these experimental therapies have made a successful transition to clinical trial and are being considered for use in human patients. The recent isolation of endothelial and cardiomyocyte precursor cells from adult bone marrow may permit the design of strategies for repair of the damaged heart. Cell-based therapies may have potential application in neovascularization and regeneration of ischemic and infarcted myocardium, in blood vessel reconstruction, and in bioengineering of artificial organs and prostheses. We expect that advances in the field will lead to the development of safer and more efficient vectors. The advent of genomic screening technology should allow the identification of novel therapeutic targets and facilitate the detection of disease-causing polymorphisms that may lead to the design of individualized gene and cell-based therapies. Melo, L. G., Pachori, A. S., Kong, D., Gnecchi, M., Wang, K., Pratt, R. E., Dzau, V. J. Gene and cell-based therapies for heart disease. Gene and cell-based therapies for heart disease. FASEB J. 18, 648 663 (2004) Key Words: coronary artery disease AAV myocardial protection contractile function Despite significant therapeutic advances, acute myocardial infarction (MI) due to coronary artery disease (CAD) remains the most prevalent cause of premature death (1, 2). The complexity of the pathological processes leading to CAD and the lack of specific predictive markers have been major impediments to the development of effective preventive therapies despite the identification of various risk factors and sensitive risk assessment technologies (3, 4). Consequently, the focus has been on the design of rescue treatments for overt symptoms of the disease such as myocardial ischemia, left ventricular pump failure, and hemodynamic overload (5, 6). Ironically, the improved survival of MI patients treated with these therapies has led directly to a dramatic increase in the number of patients suffering from heart failure (1, 2, 7), indicating the need for more effective therapies. The availability of cardiotropic vector systems such as adeno-associated virus (AAV) capable of sustained expression of therapeutic proteins (8), together with the identification of new targets for therapeutic intervention (9) and the recent isolation of bone marrowderived progenitor cells with regenerative potential (10), offers opportunities for the design of gene therapies for myocardial protection and rescue. Delivery of antioxidant, proangiogenic and contractility-enhancing genes may have potential as therapy for patients afflicted or at risk of developing MI and HF (11 13), whereas transplantation of autologous progenitor cells may have therapeutic value in revascularization and repair of ischemic and infarcted myocardium (14). Evaluation of these experimental gene and cell-based strategies for myocardial protection, rescue, and regeneration are being intensely pursued by various groups, and several small-scale trials support the feasibility of these approaches. In this article we review the major advances in gene and cell-based therapies for heart disease, with emphasis on strategies for protection and rescue of the ischemic and failing myocardium, their clinical feasibility, and a perspective on future developments in the field. Strategies for genetic manipulation of the myocardium A variety of vectors and delivery strategies are available for genetic manipulation of the myocardium with variable degrees of efficiency (Fig. 1) (15 19). The most common somatic gene therapy strategy for the myocardium involves the exogenous delivery and expression of genes whose endogenous activity may be either defective or attenuated due to a mutation or a pathological process, in order to restore the function of the deficient 1 Correspondence: Department of Physiology, Queen s University, Kingston, Ontario K7L 3N6, Canada. E-mail: melol@post.queensu.ca doi: 10.1096/fj.03-1171rev 648 0892-6638/04/0018-0648 FASEB

Figure 1. Strategies for genetic manipulation in the cardiovascular system. A) Gene transfer involves the delivery of one or several exogenous genes (transgenes) by a vector capable of expressing the therapeutic protein in the host cells. The overall goal is to increase the activity of a gene(s) (gain-of-function) whose endogenous function may be deficient and cause disease. B) Gene blockade involves inhibition of genes whose over activity may lead to disease. Two strategies are commonly used to inhibit gene activity at the transcriptional or transnational level. Short single-stranded deoxyoligonucleotides complementary to the target gene mrna (antisense oligonucleotides) are delivered to the target cells or tissue by transfection or with the aid of a vector. The antisense deoxyoligonucleotide binds to the target mrna transcript and prevents it from being translated. The second strategy uses double-stranded deoxyoligonucleotides containing the consensus binding sequences (decoy oligonucleotides) for transcriptional factors involved in the activation of pathogenic genes. Transfection of a molar excess of the decoy oligonucleotide prevents the binding and trans-activation of the genes regulated by the target transcriptional factor. Less commonly, short segments of RNA with enzymatic activity (ribozymes) are used to degrade target mrna transcripts. or undercompensating gene. In this setting, a fulllength or partial cdna encoding the deficient gene is delivered to the target tissues using a vector system capable of expressing the therapeutic protein. Such gain-of-function gene transfer strategies have been widely used with a variety of therapeutic genes, including proangiogenic and survival factors (20, 21), antioxidant enzymes (12) and anti-inflammatory cytokines (22). In other instances, the silencing of genes involved in the pathological processes may be desirable. Acute blockade of gene transcription can be achieved by treatment with short single-stranded antisense oligodeoxynucleotides, ribozymes, and, more recently, using RNA interference technology. (16, 17, 23 25). These molecules inhibit the synthesis of proteins by hybridizing in a sequence-specific complementary fashion to target mrna. Inhibition of transcription factor DNA binding using double-stranded decoy oligonucleotides containing DNA consensus binding sequences for target transcriptional factors has also been used (for a review, see ref 16). The decoy is usually delivered in molar excess, sequestering the target transcription factor and rendering it incapable of binding to the promoter region of the target gene. In many instances, short-term inhibition of a pathogenic gene is sufficient to prevent the development or progression of disease. For example, the inhibition of cell cycle regulatory proteins using decoy oligonucleotides was shown to prevent neointimal hyperplasia and subsequent restenosis after balloon angioplasty or bypass grafting (16). Recently, nucleic acid and peptide aptamers have been developed that are capable of inhibiting protein function without altering the genetic complement of the host (26). However, the use of aptamers in cardiovascular therapeutics has not been evaluated. The choice of therapeutic target, vector and delivery strategy is governed to a large extent by the pathological features of the disease, the putative role of the target gene(s) in the pathophysiological process, and the timing of intervention (15). The efficiency of gene transfer to the myocardium is highly dependent on the type of vector, route, and the dosage and volume of delivery of the genetic material. (27, 28). A number of vector systems have evolved over the years (Table 1). Nonviral vectors, which include naked plasmids, cationic liposome formulations, synthetic peptides, and several physical methods, usually yield low and transient gene transfer efficiency due to lack of genomic integration and rapid degradation of the vector (19). A promising new delivery strategy uses synthetic peptide carriers containing a nuclear localization signal to facilitate nuclear uptake of the target cdna (29). These peptide DNA heteroplexes are recognized by intracellular receptor proteins and imported into the nucleus, where the target cdna is transcribed. Recombinant viruses have become the preferred vectors for myocardial gene transfer (Table 1). These are replication-deficient viral particles that retain their ability to penetrate target cells and deliver genetic material with much higher efficiency than nonviral vectors (18, 30). Some vectors, such as AAV and lentivirus, are capable of sustained expression of the thera- MOLECULAR THERAPIES FOR HEART DISEASE 649

TABLE 1. Vectors used for transfer and manipulation of genetic material in cardiovascular tissues Vector Chromosomal integration Transfer efficiency in vivo Onset of transgene expression Sustainability of therapeutic effect Level of expression Target cells Host immune response Potential risks Nonviral Cationic liposomes No rapid short quiescent and cytotoxicity dividing HVJ-liposomes No rapid short quiescent and cytotoxicity dividing Naked plasmid No moderate short quiescent and cytotoxicity dividing Viral Retrovirus Yes rapid life-long dividing cytotoxicity oncogenesis Lentivirus Yes rapid life-long quiescent and dividing Adenovirus No rapid moderate quiescent and dividing Adeno-associated virus Yes slow life-long quiescent and dividing Herpes simplex virus No moderate long quiescent and dividing Alphavirus No very rapid short quiescent and dividing cytotoxicity viral mutation cytotoxicity viral mutation oncogenesis viral mutation cytotoxicity viral mutation cytotoxicity viral mutation peutic gene (8, 18, 30), which may be essential for the design of therapies for chronic myocardial disease. Unfortunately, a robust immune reaction may be triggered by the host in response to the viral proteins synthesized by viral vectors that may reduce the efficiency of gene transfer and the sustainability of transgene expression (18). There is a risk, albeit remote, that these vectors may revert to replication proficiency, thus raising safety concerns about biological hazards such as oncogenesis and insertional mutagenesis (18, 30). The overall safety and specificity of gene transfer protocols could, however, be enhanced by incorporating regulatory elements capable of directing tissuespecific expression as well as regulated expression of the transgene in response to underlying pathophysiological cues such as hypoxia, oxidative stress, or inflammation (31). Several routes have been used to administrate the therapeutic material to the myocardium (Fig. 2). Intracoronary delivery of the therapeutic gene is the preferred route for global myocardial diseases such as heart failure and cardiomyopathy. However, the selectivity of coronary endothelium and the barrier imposed by the basement membrane may restrict the diffusion of some vectors and limit distribution and uptake of the therapeutic transgene. In contrast, intramyocardial injection may be a desirable method for gene delivery to areas of regional myocardial disease. This approach has been used to deliver angiogenic and cytoprotective genes to ischemic myocardium (12, 15). However, transgene expression is restricted to the area surrounding the site of injection, requiring in some cases multiple injections to adequately cover the affected area. A variety of catheter types have been developed for both intracoronary and intramyocardial delivery with the assistance of trans-esophageal echocardiographic guiding and mapping techniques (15, 32). Other methods, such as pericardial injection and retroperfusion, have had limited application in myocardial gene transfer (33). Targets for gene therapy for myocardial protection and rescue Several genes have emerged as potential targets for gene therapy for myocardial disease (Table 2). In the setting of myocardial protection, the overexpression of cytoprotective and survival genes such as antioxidant enzymes (11, 34, 35), antiapoptotic proteins (36), protein kinase B/Akt (21), and/or the inhibition of proinflammatory cytokines (22), proapoptotic (37), and prooxidant (38) genes have emerged as potential therapeutic targets for cardioprotection from studies in various animal and cellular models of myocardial ischemic injury. Gene therapy strategies for rescuing failing myocardium may be attainable in certain conditions (Table 2). Therapeutic angiogenesis by delivery of genes coding for proangiogenic growth factors has been shown to promote neovascularization and functional recovery of ischemic myocardium in several animal models and in humans 650 Vol. 18 April 2004 The FASEB Journal MELO ET AL.

Figure 2. Strategies for delivery of therapeutic genes to the myocardium. Five methods have been used to deliver genetic material to the myocardium. Epicardial an intramyocardial injection deliver genetic material locally to the epicardial surface and the myocardium, respectively. Intramyocardial delivery can also be achieved by endocardial injection with a specialized intraventricular catheter. In all cases, the concentration of injected material decreases with distance from the site of injection. Intracoronary infusion with the aid of a catheter distributes the therapeutic agent to the regional field supplied by the source coronary artery. Delivery of the genetic material to the aortic root or to the left ventricle cavity (intraventricular delivery) concurrent with aortic cross-clamping allows global myocardial distribution of the injected material by retrograde perfusion. A similar effect can be achieved by retroinfusion of the genetic material into the coronary sinus or cardiac vein. with coronary artery disease (12, 15, 20, 39 41). Strategies in the postinfarction period may include inhibition of genes involved in regulation of ventricular remodeling and chamber dilatation such as the matrix metalloproteinases (MMPs) that participate in extracellular matrix degradation (42). Potential strategies for rescuing contractile function in the failing myocardium include overexpression of the sarcoplasmic reticulum calcium ATPase (SERCA2a) (43), -adrenergic receptor (44), and adenylate cyclase (45) (Table 2). An exciting new field is emerging with the recent identification and isolation of endothelial and cardiomyocyte precursor stem cells from adult bone marrow (46); the ability to expand and genetically modify these cells ex vivo offers the opportunity to use them as an autologous cellular substrate for repair of infarcted myocardium. Gene therapy for myocardial protection Myocardial ischemia due to CAD initiates a continuum of myocardial injury that is perpetuated by reperfusion (I/R injury) (47) (Fig. 3). Reoxygenation of the ischemic myocardium increases the formation of reactive oxygen species (ROS) formation, which may eventually deplete the buffering capacity of endogenous antioxidant systems (48). The development of gene therapies for acute myocardial infarction has been difficult because the time required for transcription and translation of therapeutic genes with the current generation of vectors exceeds the window for successful intervention. For this reason, gene transfer of anticoagulant genes is not feasible as primary thrombolytic therapy for acute myocardial infarction. An alternative gene therapy approach for myocardial protection is to devise a strategy that could prevent I/R injury by using a method that could confer long-term expression of cytoprotective genes in the myocardium (Table 2). This novel concept of preventive gene therapy would protect the heart from future I/R injury, thereby minimizing the need for acute intervention (11). Gene therapy aimed at increasing endogenous antioxidant reserves should, in principle, be a useful strategy given the prominent role of oxidative stress in CAD and I/R injury (47). Such a strategy may potentiate the native protective response of the myocardium, rendering it resistant to future ischemic insults. We have evaluated the feasibility of antioxidant enzyme gene transfer as a long-term first line of defense against I/R-induced oxidative injury using an raav vector for intramyocardial delivery of heme oxygenase-1 (HO-1), an enzyme that is involved in the catabolism of heme (12). Our findings show that HO-1 gene delivery to the left ventricular risk area several weeks in advance of myocardial infarction in a rat model of myocardial I/R injury results in 80% reduction in infarct size (Fig. 4). The reduction in myocardial injury in the treated animals is accompanied by decreases in oxidative stress, inflammation, and interstitial fibrosis, leading to postinfarction functional recovery and normalization of left ventricular dimensions (Fig. 4). Comparable findings were seen with extracellular superoxide dismutase (ecsod) gene transfer (36, 49). This secreted metalloenzyme plays an essential role in maintenance of redox homeostasis by dismutating the oxygen free radical superoxide. Our findings showed improved long-term survival after MI in the ecsod-treated animals relative to the animals treated with the control vector in parallel with smaller infarcts and decreased myocardial inflammation (49). Significant protection from I/R injury has also been achieved by overexpression of other major antioxidant enzyme systems such as Cu/Zn SOD (50), catalase (51), and glutathione peroxidase (52), stress-induced heat shock proteins (53), survival genes (Bcl-2, Akt) (21, 54), as well as immunosuppressive cytokines (22), adenosine A 1 and A 3 receptors (55), kallikrein (56), caspase inhibitor (37), and hepatocyte growth factor (57). Thus, the long-term overexpression of anti-oxidant enzymes and cytoprotective genes in the myocardium provides a strategy for enhanced protection from I/R-induced injury. A potential strategy for acute protection from I/R injury may be the inhibition of proinflammatory genes activated by I/R. Morishita et al. (58) showed that pretreatment with a decoy oligonucleotide capable of MOLECULAR THERAPIES FOR HEART DISEASE 651

TABLE 2. Targets for gene-based therapy for acquired, inherited, and congenital heart disease a Strategy Therapeutic target Genetic manipulation Vector Application Protection/ prevention Rescue Inherited heart disease Congenital heart disease Antioxidant enzymes HO-1, SOD, catalase, GPx overexpression AAV, LV CAD, MI Heat shock proteins HSP70, HSP90, HSP27 overexpression AAV, LV CAD, MI Anti-inflammatory I-CAM, V-CAM, NF- B, TNF- inhibition AS-ODN Decoy ODN AAV-AS-ODN RV-AS-ODN graft atherosclerosis transplantation Survival genes Bcl-2, Akt overexpression AAV, LV CAD, MI, HF Proapoptotic genes Bad, p53, Fas ligand inhibition AS-ODN Decoy ODN AAV-AS-ODN MI, HF Coronary vessel tone enos, adenosine (P1, P3) receptors overexpression RV, AAV(?) CAD, HF Proangiogenic genes VEGF, FGF, HGF, Ang-1, MCP-1 overexpression AAV CAD, MI, HF IGF-1, HIF-1 /VP16, Prox-1 Contractility/calcium regulation -Adrenergic receptors, SERCA 2A, overexpression AAV HF V 1 receptor BARK, phosphalamban Inhibition AAV HF Plaque stabilization CD40 overexpression RV, AAV(?) CAD Channelopathies SCN5A, I k overexpression/ inhibition -MHC-AAV arrhythmia Cardiomyopathy Sarcomeric proteins, sarcoglycans overexpression -MHC-AAV DCM Heart and vessel defects Endoglin, NKx2.5, TBX5, TFAP2B overexpression -MHC-AAV Septal defects, patent ductus arteriosus, arteriovenous malformations, looping conductance defects a AAV, adeno-associated virus; AS-ODN, antisense oligodeoxynucleotide; CAD, coronary artery disease; DCM, dilated cardiomyopathy; HF, heart failure; LV, lentivirus; MI, myocardial infarction, -MHC, alpha myosin heavy chain; RV, retrovirus; HO-1, heme oxygenase-1; SOD, superoxide dismutase; GPx, glutathione peroxidase; HSP70, 70 KD heat shock protein; HSP90, 90 KD heat shock protein; I-CAM, intracellular adhesion molecule; V-CAM, vascular adhesion molecule; NF- B, nuclear factor kappa B; TNF-, tumor necrosis factor alpha; enos, endothelial nitric oxide synthase; VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; HGF, hematopoietic growth factor; SERCA2A, sarcoplasmic/endoplasmic reticulum Ca 2 ATPase; V1, vasopressin-1 receptor; ARK, beta adrenergic receptor kinase; SCN5A, cardiac sodium channel gene 5A. inhibiting the trans-activating activity of NF- B reduces myocardial infarct after coronary artery ligation in rats. Although the rapid in vivo degradation of oligonucleotides precludes their use in long-term myocardial protection, this strategy may be useful in treatment of acute myocardial ischemia in cardiac transplantation (59) For example, treatment with antisense oligonucleotide directed against intercellular adhesion molecule-1 (ICAM-1) prolongs cardiac allograft tolerance and long-term survival when administered ex vivo before transplantation into the host (60). Such an approach could be useful in the preparation of donor hearts for transplantation. Gene therapy for myocardial ischemia Therapeutic angiogenesis by exogenous supplementation of proangiogenic factors has emerged as a potential treatment option for patients for whom percutane- 652 Vol. 18 April 2004 The FASEB Journal MELO ET AL.

Figure 3. Pathophysiology of coronary artery disease. Myocardial ischemia subsequent to coronary artery occlusion, if sufficiently prolonged will result in irreversible cellular damage highlighted by alterations in membrane fluidity and pump activity, mitochondrial damage and depressed metabolic and contractile activity. Reperfusion introduces a separate set of cellular stresses that may exacerbate the damage initiated during ischemia. Reoxygenation of the ischemic myocardium results in the formation of reactive free radical oxygen species (ROS), leading to activation of the inflammatory cascade, myocyte injury and endothelial dysfunction. The end result of I/R is a continuum of myocardial injury that culminates with myocardial infarction caused by membrane damage, contractile dysfunction, and eventual cell death, In time these changes lead to ventricular remodeling, a process characterized by myocyte hypertrophy, interstitial fibrosis, chamber dilatation, and increased propensity for contractile dysfunction and failure. Acute coronary syndromes are a clinical manifestation of coronary artery disease. Chronic coronary artery disease leads to ischemic cardiomyopathy heart failure and premature death. ous angioplasty or surgical revascularization has been excluded. Evidence of enhanced neovascularization and functional recovery has been demonstrated in several animal models of hindlimb and myocardial ischemia by gene transfer of VEGF (20, 61 63; for review, see refs 12, 15), FGF (40) and hepatocyte growth factor (HGF) (41). In all cases, improvement in tissue perfusion was accompanied by morphological and angiographic evidence of new vessel formation, establishing a relationship between improved tissue viability and neovascularization. For example, Mack et al. (39) showed improvement in regional myocardial perfusion and left ventricular function in response to stress after intramyocardial delivery of VEGF 121 by adenovirus in an ameroid constrictor model of chronic myocardial ischemia in pigs. Using a similar model, Giordano et al. (40) showed improvement in blood flow and a reduction in stress-induced functional abnormalities after intracoronary injection of an adenovirus vector encoding human FGF-5 as early as 2 wk after placement of the ameroid in the proximal left circumflex artery. The improved blood Figure 4. Gene therapy strategy for long-term myocardial protection. Effective, long-lasting gene therapy for myocardial protection can be achieved by overexpressing cardioprotective genes with a vector such as adeno-associated virus (AAV) capable of conferring long-term expression of the therapeutic gene. In this example, we demonstrate the effectiveness of raav-mediated transfer of the gene coding for the cytoprotective enzyme heme oxygenase-1 (HO-1) in conferring long-term myocardial protection from I/R injury. A) The therapeutic gene (HO-1) or reporter gene (LacZ) was delivered by intramyocardial injection to the left ventricle territory supplied by the left anterior descending coronary artery (LAD). B, C) Eight wk after gene transfer, I/R injury was induced by ligation and release of the LAD. Gross (TTC) an microscopic (H&E) histological analysis of the infarcted region 24 h after reperfusion revealed significant reduction in myocardial injury D) Echocardiographic analysis of left ventricular function and chamber dimensions 1 month after acute myocardial infarction showed normalization of ventricular function (fractional shortening, FS; ejection fraction, EF) and left ventricle dimensions [left ventricular diastolic (LVDD) and systolic (LVSD) diameters] in the HO-1-treated animal. MOLECULAR THERAPIES FOR HEART DISEASE 653

flow was attributed to an increase in capillary to fiber ratios. The success of these preclinical studies led to several small-scale nonrandomized phase I and II clinical trials with patients suffering from myocardial ischemia (20, 61, 62, 64 65; for a review, see ref 66). These safety trials, although consisting of small nonrandomized patient samples have demonstrated the potential of angiogenic gene therapy for treatment of ischemic heart disease. Losordo et al. (20) carried out a phase I study in five male patients aged 53 71 years of age with CAD who did not respond to conventional anti-anginal therapy. The authors reported that direct intramyocardial delivery of naked plasmid encoding VEGF 165 into the ischemic myocardium led to reduction of anginal symptoms and improvement, albeit modest, in left ventricular function concomitant with reduced ischemia. Rosengart and colleagues (61) reported significant improvement in regional ventricular function and wall motion in the region of vector administration after intramyocardial delivery of VEGF 121 in patients undergoing conventional coronary artery bypass grafting compared with patients receiving placebo. Using catheter-based delivery of naked VEGF 165 assisted by electromechanical NOGA mapping of the left ventricle in patients with chronic myocardial ischemia, Vale and colleagues (64) reported significant reductions in weekly anginal attacks for as long as 1 year after gene delivery in the treated patients compared the patients receiving placebo. Recently, the results of the angiogenic gene therapy (AGENT) double-blinded, randomized, placebo-controlled trial using dose-escalating adenovirus-mediated intracoronary delivery of FGF-4 in patients with angina showed increased exercise tolerance and improved stress echocardiograms 4 and 12 wk after gene transfer in the patients that received FGF-4 gene therapy compared with patients receiving placebo (65). Unfortunately the outcome beyond 12 wk has not been reported. Despite the promising findings of these small-scale trials, there are several outstanding issues relating to the safety and sustainability of the approach. The safety of therapeutic angiogenesis requires systematic evaluation. This is significant in light of recent evidence that constitutive overexpression of VEGF in mouse heart led to intramural angiomas, followed by heart failure and death (67), and may accelerate plaque progression in atherosclerotic vessels (68). This observation underscores the necessity for regulated expression of proangiogenic factors. Such a strategy may require the incorporation of promoter sequences such as hypoxiasensitive responsive elements capable of rendering expression of the therapeutic transgene subservient to the pathophysiological changes in myocardial oxygen tension. This concept has been validated by Su et al. (69), who demonstrated that VEGF expression by an AAV vector in ischemic myocardium could be regulated by hypoxia by incorporation of the erythropoietin responsive element (HRE). Another approach to achieve regulated therapeutic angiogenesis uses engineered transcription factors capable of activating endogenous VEGF expression as a strategy to induce VEGF expression in pathophysiological conditions (70). These novel strategies may allow endogenous regulation of angiogenesis so that the magnitude of neovascularization is matched to the severity of the ischemic insult. More work is necessary to determine the safest and most efficacious route and method of therapeutic gene delivery in order to avert potentially hazardous late onset side effects such as neovascularization of occult neoplasms or peripheral vascular effects that may result in edema and hypotension. The optimal strategy may require targeted tissue delivery by incorporation of cell-specific promoters for expression of the transgene exclusively at the target sites. Regarding the therapeutic sustainability, it is necessary to establish whether the desired long-term therapeutic effect can be achieved with a single administration of the therapeutic gene or whether multiple treatments may be required. This is an important aspect because VEGF-induced neovessels tend to regress soon after termination of transgene expression in the absence of adequate blood supply (71). Gene therapy for rescue of contractile function Rescue of contractile function in the failing myocardium is another potential target for gene therapy. The failing myocardium is characterized by alterations in calcium handling, decreased myofilament sensitivity, excessive catecholamine release, and adrenergic receptor down-regulation and desensitization (72), leading to decreased contractility. -Adrenergic receptors ( - AR) are G-protein-coupled receptors that play an essential role in regulation of myocardial contractility and inotropic state in response to neurohumoral stimulation (73). The -AR signaling and calcium regulating pathways have for several years been used as targets for treatment of heart failure (72). Recent preclinical studies suggest that genetic manipulation of these therapeutic targets may be a viable strategy for treatment of heart failure (Table 2). Adenovirus-mediated intracoronary delivery of the 2 -AR gene led to improvements in basal and isoproterenol-stimulated LV contractility and hemodynamic function in rabbits (44), and rescued -AR signaling in ventricular myocytes from failing hearts (13, 74). Similarly, adenovirus delivery of the -ARKct peptide inhibitor improved postinfarction LV function significantly in rabbits after myocardial infarction in parallel with increased -ARmediated stimulation of adenylate cyclase activity and camp generation (75). Recently, Roth et al. (45, 76) demonstrated that cardiac-specific overexpression of adenylate cyclase type VI (AC VI ) improves ventricular function, restoring -AR-stimulated camp generation and increasing long-term survival in mice rendered cardiomyopathic by overexpression of Gq protein. These findings suggest that gene transfer protocols aimed at normalizing -AR signaling may have applica- 654 Vol. 18 April 2004 The FASEB Journal MELO ET AL.

tion as a therapeutic strategy for functional rescue of the failing heart. For example, exogenous overexpression of -AR receptors and signaling proteins by gene transfer could be used as a strategy to compensate for the decrease in endogenous -AR density and sensitivity resulting from chronic sympathetic activation in heart failure, thereby normalizing left ventricular function. Gene therapy strategies for normalization of myocardial cytosolic calcium transients have also shown promising results in experimental models of heart failure (Table 2) (43, 77, 78; for a review, see ref 79). The ratio of phospholamban to SERCA2a is increased in heart failure, resulting in decreased Ca 2 ATPase activity and reduced calcium uptake by the SR (72). Adenovirusmediated overexpression of SERCA2a in neonatal cardiac myocytes enhances contraction by increasing peak [Ca 2 ] i release and a decrease in resting [Ca 2 ]i (78). In a rat model of heart failure induced by aortic banding, intracoronary SERCA2a gene delivery by adenovirus at approximately the time of transition from compensated hypertrophy to heart failure restored systolic and diastolic function concomitant with an increase in basal Ca 2 -ATPase activity (44), and improved phosphocreatine/atp ratio and long-term survival (77). SERCA2a gene transfer normalized cytosolic transients and restored contractile function in ventricular myocytes isolated from patients with end-stage heart failure (80). Presumably, overexpression of SERCA2a restores the normal stoichiometry between phospholamban and the Ca 2 -ATPase, thereby preventing cytosolic calcium overload and left ventricular dysfunction. Conversely, antisense inhibition of phospholamban was shown to improve contractility in ventricular myocytes of end-stage heart failure patients (81) in association with improved calcium sensitivity of SERCA and reduced time for recovery of the Ca 2 transient. The long-term efficacy and safety of adenoviral-mediated myocardial expression of adrenergic and calcium-regulating proteins remain to be established. Sustained expression of the therapeutic transgene may be essential for rescue of the failing heart, thus necessitating the use of a vector type capable on long-term transgene expression. Second, the physiological consequences of chronic -AR and SERCA2a overexpression are not known. Although transgenic mice with cardiacspecific overexpression of 2 -AR or SERCA2a do not show any morphological evidence of myocardial pathology, it is not known whether viral-mediated expression of these proteins has any secondary effects besides calcium regulation and inotropic state. Concerns have been raised that the increase in SERCA2a expression by gene transfer in the failing heart may impose extra demands on myocardial energy expenditure due to increased inotropic state and may cause adverse electrophysiological events such as arrhythmias (for a review, see ref 82). Such potential adverse effects could accelerate myocardial cell death and precipitate the progression of heart failure, and will have to be addressed before inotropic gene therapy could be considered for clinical trial. Gene therapy for inherited and congenital heart disease In principle, myocardial disease resulting from single gene mutations could be corrected by exogenous delivery of the normal gene. However, the unavailability of vectors capable of efficient long-term gene expression has been a major drawback in the design of rescue therapies for inherited heart disease. Preclinical data supports the feasibility of gene therapy for some forms of inherited cardiomyopathy (Table 2). Kawada et al. (82) showed that intramyocardial delivery of -sarcoglycan to 5-wk-old TO-2 Syrian hamsters using an AAV vector completely rescued the progression of cardiomyopathy and led to a drastic increase in life expectancy. Similarly, Ikeda et al. (83) showed that coronary retroinfusion of adenovirus vector coding for -sarcoglycan in 8- to 12-wk-old BIO 14.6 hamsters, resulted in restoration of -, -, and -sarcoglycan to the sarcolemma and improved ventricular function compared with age-matched untreated CM hamsters. Myocardial delivery of genes encoding defective channel proteins or regulatory G-proteins may provide a strategy for correction of the genetic defects associated with inherited and acquired LQT syndromes. Nuss et al. (84) reported that adenoviral transfer of the human HERG gene to adult rabbit ventricular myocytes maintained in primary culture led to abbreviated action potentials and drastically reduced the incidence of early afterdepolarizations after a train of action potentials. This was found to be associated with increased duration of the refractory period. Similarly, Donahue et al. (86) were able to reduce heart rate after atrial fibrillation in pigs by local delivery of G i2 gene to the atrioventricular node by adenovirus, suggesting this approach may have application in the treatment of atrial arrhythmias. The design of gene therapies for congenital heart and vessel disease is problematic. The precision, both in time and mechanism, by which these developmentally regulated genes exert their effects on heart morphogenesis and development (87, 88) dictates that an external corrective measure such as replacement of defective genes, needs to be performed before the developmental programs affected by the mutated genes are activated, since the anatomical and functional defects emanating from these mutations may be irreversible (88). Second, the dependency of normal heart and vessel development and maturation on precise stagespecific regulation of these morphogenetic genes (87) mandates that corrective strategies be amenable to regulation by the endogenous mechanisms responsible for normal development. The ability to intervene and reprogram a defective gene within the crucial developmental time window requires the availability of diagnostic tools that would permit detection of such mutations before the onset of disease and access to an effective MOLECULAR THERAPIES FOR HEART DISEASE 655

system for in utero gene delivery. The prohibitive costs of screening technologies for congenital diseases currently restrict its use to cases with a strong familial history, leaving many cases undiagnosed. Cell-based therapy for myocardial rescue and repair An alternative strategy for neovascularization of the ischemic heart involves the use of endothelial progenitor cells (EPC) as angiogenic substrate. These bloodborne cells are thought to originate from a common hemangioblast precursor in adult bone marrow (10, 89; for a review, see ref 14), express endothelial lineage markers (i.e., CD34, Flk-1, VE-cadherin, PECAM-1, von Willebrand factor, enos, and E-selectin) (for review, see refs 14, 90), and can be expanded and genetically modified ex vivo to yield sufficient numbers for therapeutic applications (Fig. 5) (10, 91 93). The cells, whose abundance is low in basal conditions are recruited to sites of injury, such as ischemic myocardium and damaged blood vessels, where they may participate in local vasculogenesis and repair of the injured vessels (90, 94, 95). Exogenous administration of cytokines such as VEGF and G-CSF increases the numbers of circulating EPC several-fold by mobilizing them from the bone marrow (10, 90, 96 98). The therapeutic potential of EPC for neovascularization and rescue of ischemic myocardium has been demonstrated. Myocardial transplantation of autologous CD31 cells isolated from peripheral blood induced new vessel formation and improved left ventricular perfusion and performance in pig hearts rendered ischemic by placement of an ameroid constrictor in the circumflex coronary artery (99). Likewise, transplantation of whole (93) or CD34 -selected human mononuclear cells from peripheral blood immediately after acute myocardial infarction in nude rats led to revascularization and repair of the infarcted myocardium, resulting in reduced interstitial fibrosis and improved ventricular function. Favorable results have also been seen with bone marrow-derived mononuclear cells (BM-MNC). Kocher et al. (100) reported that intravenous delivery of human CD34 BM-MNC to nude rats with myocardial infarction led to significant neovascularization of the infarcted myocardium, resulting in reduced apoptosis of myocytes in the peri-infarct area, decreased fibrosis and sustained recovery of left ventricular function. Others have reported that implantation of bone marrow-derived Lin - c - kit (101) or side population (SP) cells (102) in the infarct border resulted in improved left ventricular function in association with new vessel formation. In addition to EPC, various other cell sources, such as skeletal myoblasts, embryonic and fetal cardiomyoctes, and bone marrow-derived myocyte progenitors have been used for cellular cardiomyoplasty of infarcted myocardium (Table 3) (for review, see refs 14, 103), but the efficacy of these cells in repair myocardium has been inconsistent. The use of an adult self-regenerating autologous source of progenitor cells with the potential for differentiating into cardiomyocytes would appear to be ideal. Mesenchymal cells from the bone marrow stroma of long bones (MSC) may offer a viable option for cellular cardiomyoplasty using autologous cells. These cells exhibit a high degree of plasticity (104) and can differentiate into functional cardiomyocytes under specific culture conditions (102, 104, 105). MSC can be induced to differentiate into synchronously beating Figure 5. Isolation, cultivation, and genetic engineering of endothelial progenitor cells (EPC) for therapeutic application. EPC can be isolated from the mononuclear cell fraction of bone marrow, peripheral blood, or umbilical chord blood with or without further selection and purification. Mononuclear cells are expanded ex vivo under endothelial-specific growth conditions and may be genetically modified to overexpress one or several therapeutic genes. The differentiated cells are then used in transplantation protocols for rescue and repair of damaged tissues such as infarcted myocardium, ischemic limb or damaged muscle. The cells may also be used for endothelialization of damaged blood vessels and vascular prosthetic grafts and in tissue engineering. 656 Vol. 18 April 2004 The FASEB Journal MELO ET AL.

TABLE 3. Substrates for cell-based therapies for the heart a Cell type Source Advantages Disadvantages Applications Skeletal myoblasts (satellite cells) Fetal/neonatal cardiomyocytes Embryonic stem cells Bone marrow stromal cells (mesenchymal stem cells) Endothelial progenitor cells (EPC) Autologous skeletal muscle biopsy Allogenic fetal/ neonatal heart Blastocyst (inner mass) Bone marrow Bone marrow Peripheral blood Chord blood Ease of isolation and culture scalability Genetic engineering Autologous source (no rejection) ischemiaresistant Ability to establish E-C coupling High proliferative rate genetic engineering Pluripotent Ability to establish E-C coupling Genetic engineering Multipotent Ease of isolation and culture Autologous source (no rejection) Genetic engineering Ease of isolation and culture scalability Autologous source (no rejection) Genetic engineering Inability to establish E-C coupling Potential arrhythmias Uncontrolled proliferation Controversial ethical/moral issues Donor availability Poor long-term survivability Immune rejection Sensitive to ischemic injury Controversial ethical/moral issues Donor availability Large-scale expansion difficult Allogenic immune rejection Dependence on murine feeder layers Insufficient characterization Insufficient characterization scalability Insufficient characterization MI HF Cardiomyopathy MI HF Cardiomyopathy Long QT disorders MI HF Cardiomyopathy Long QT disorders Tissue engineering Organogenesis MI HF Cardiomyopathy Tissue engineering Organogenesis CAD HF Blood vessel repair Vascular protheses Tissue engineering a CAD, coronary artery disease, E-C, electromechanical coupling; EPC, endothelial progenitor cell; MI, myocardial infarction; HF, heart failure. cardiomyocytes in vitro after treatment of primary cultures of mouse bone marrow with the cytosine analog 5-azacytidine (106, 107). Several groups have provided evidence that the bone marrow mesenchymal cell population may contain cardiomyocyte precursors. Toma and colleagues (108) showed that human MSC transplanted into the left ventricular wall of immunodeficient mice differentiate into cardiac myocytes without the need for myogenic differentiation before transplantation. Tomita et al. (107) reported that transplantation of 4-azacytidine-treated bone marrow cells repopulate the scar and significantly improve left ventricular function in cryoinjured rat hearts. Wang et al. (109) detected several cell types, including cardiomyocytes, endothelial cells, and fibroblasts, within and on the border of the scar 1 month after intracoronary delivery of retrovirally transduced isogenous bone marrow cells to infarcted rat hearts, suggesting that factors emanating from the injured myocardium may induce trans-differentiation of bone marrow progenitors into the various cell types required for regeneration and maintenance of the myocardium. More recently, several groups reported evidence of extracardiac progenitors in necropsy specimens of hearts obtained from subjects that had undergone sex-mismatched heart (110, 111) or bone marrow transplantation (112). Quaini et al. (110) and Muller et al. (111) reported the presence of highly proliferating Y chromosome-positive myocytes and vascular cells in myocardial specimens from male recipients that had received hearts from female donors. The recipientderived cells expressed stem cell-related antigens, including c-kit, MDR1, and Sca-1 (110) and connected by gap junction with neighboring myocytes (111), indicating that these precursor cells develop into functional cardiomyocytes. The bone marrow is a likely source from which these extracardiac precursor cells are mobilized in response to the injury. The cells probably migrate to the injured myocardium in response to locally released cues such as cytokines, where they may home and participate in the repair of the damaged myocardium (for review, see refs 14, 113). In addition to the marrow-derived precursor cells, clusters of highly proliferating primitive cells have been detected in the infarcted myocardium by two independent groups (114 116; for a review, see ref 113). In a recently published study, Beltrami and colleagues (115) reported the identification and isolation of a population of undifferentiated lineage negative (Lin - ) cells that express stem cell markers such as c-kit (ckit Pos ) and stem cell antigen 1 (Sca-1 Pos ). These cells were found to be clonogenic and self-renewing and capable of differentiating into all myocardial cell types, including cardiomyocytes, endothelial cells, and vascular smooth muscle cells (115). Within the clusters, the progenitor cells were found to be at different stages of MOLECULAR THERAPIES FOR HEART DISEASE 657

cardiomyogenic differentiation, reflecting their cardiogenic potential. This was further supported by the ability of early passage cells to induce significant myocardial regeneration and improve ventricular performance after transplantation into infarcted hearts from syngeneic rats (115). Recently the same group reported a dramatic 13-fold increase in the abundance of these proliferating stem cells in the myocardium of patients with aortic stenosis (116), raising the intriguing specter that the compensatory cardiac enlargement triggered by the stenosis in these patients is due at least in part to hyperplastic growth induced by these proliferating cells. The identification of resident cardiac progenitors may provide an explanation for the controversial early findings by the same group of actively dividing myocytes in the ventricular myocardium of healthy and end-stage heart failure patients (117). The mitotic index in explants from patients with ischemic or idiopathic heart failure was several-fold higher than in explants from patients without heart disease, suggesting that the resident precursor cells constitute a cardiac self-repair mechanism for the replacement of damaged or dying myocytes. Such a mechanism could also potentially play a role in renewal of myocytes lost as a result of biological turnover and cellular aging in the normal heart (117). However, the regenerative capacity of this self-repair mechanism has been questioned by some groups on the basis that the small number of resident cells and extracardiac progenitors that are capable of migrating to the heart is insufficient to induce effective long-term regeneration of the myocardium (118, 119). Some controversy remains regarding the relative roles of cardiac stem cell differentiation and fusion to the regenerative process. Whereas Beltrami et al. (115) found no evidence of cell fusion after transplantation of c-kit Pos cardiac progenitors into the infarcted heart, Oh and colleagues (120) showed roughly equal contributions of differentiation and fusion to myocardial regeneration. Using a Cre-Lox donor/recipient pair of transgenic mice, this group showed that myocardial regeneration resulted equally from differentiation and fusion of the donor cells with the recipient cells after intravenous administration of Sca-1 Pos cells isolated from the hearts of -MHC-Cre into Cre-dependent, LacZ-expressing transgenic mice (R26R) with myocardial infarction. Regardless of the nature of the mechanism(s) of myocardial reconstitution, these studies nevertheless advert to the existence of bona fide resident cardiac progenitors and their potential role in cardiac cell renewal and regeneration. These encouraging preclinical results have led to several recent small-scale feasibility and safety studies to evaluate the therapeutic potential of bone marrow cell transplantation in treatment of ischemic heart disease and myocardial infarction (121 125). Strauer et al. (121) reported that intracoronary delivery of unfractionated autologous mononuclear bone marrow cells 6 days after infarction led to a reduction in infarct size and improvement in ventricular function and chamber geometry 10 wk after transplantation. In a recent small-scale phase I clinical trial, Stamm and colleagues (122) injected autologous AC133 bone marrow cells into the infarct border during CABG in 6 patients that had suffered earlier acute transmural myocardial infarction. The authors reported improved perfusion of the infarcted area and significant enhancement of global left ventricular function 3 9 months after surgery. Using a randomized group of 20 patients with reperfused acute MI, Assmus and colleagues (123) reported that intracoronary infusion of either BM-MNC or PB-MNC 4 days after infarction led to significant improvement in global left ventricular ejection fraction and wall motion in the infarct zone and reduced systolic dimensions at 4 months follow-up, in association with increased coronary flow reserve in the infarct artery and greater viability in the infarct zone. Two other groups have reported that trans-endocardial delivery of autologous BM-MNC using NOGA mapping led to significant improvements in left ventricular perfusion and performance in patients with end-stage ischemic heart disease (124) or with stable angina (125). Nevertheless, these findings should be considered preliminary. The nature of the mobilizing, migration, and homing signals for bone marrow progenitor cells and the mechanism of differentiation and incorporation into the target tissues need to be identified. Multimember controlled trials are needed in order to define and standardize the optimal time and method of delivery and effect. A significant number of the transplanted cells die or detach soon after transplantation, suggesting that strategies to improve cell survival and adhesion at the time of transplantation may be necessary. We showed recently that genetic modification of the bone marrow mesenchymal cells before grafting with a retroviral vector overexpressing the survival gene Akt1 significantly reduced peri-infarct death of the transplanted cells, leading to reduced myocardial scarring and improved ventricular function (126). Finally, the morphological and functional complexity of the myocardium raises a cautionary point against designing overly simplistic grafting protocols and suggests that the optimal grafting procedure for cardiac repair may require more than one cell typeofor example, cardiomyocytes, fibroblasts, and endothelial cellsoto produce a graft that is able to recapitulate normal cardiac function. Tissue engineering An emerging area where cell transplantation and genetic manipulation may play a pivotal role is in the bioengineering of artificial organs and prostheses (127). Advancement in this area is conditioned by the development of immunocompatible biodegradable materials that could be used as scaffolds shaped to the desired configurations and to which autologous cells may be seeded to provide a biologically active surface. 658 Vol. 18 April 2004 The FASEB Journal MELO ET AL.