Unlocking the Pragmatic Potential of Regenerative Therapies in Heart Failure with Next-Generation Treatments

Patients with chronic heart failure (HF) have a poor prognosis due to irreversible impairment of left ventricular function, with 5-year survival rates <60%. Despite advances in conventional medicines for HF, prognosis remains poor, and there is a need to improve treatment further. Cell-based therapies to restore the myocardium offer a pragmatic approach that provides hope for the treatment of HF. Although first-generation cell-based therapies using multipotent cells (bone marrow-derived mononuclear cells, mesenchymal stem cells, adipose-derived regenerative cells, and c-kit-positive cardiac cells) demonstrated safety in preclinical models of HF, poor engraftment rates, and a limited ability to form mature cardiomyocytes (CMs) and to couple electrically with existing CMs, meant that improvements in cardiac function in double-blind clinical trials were limited and largely attributable to paracrine effects. The next generation of stem cell therapies uses CMs derived from human embryonic stem cells or, increasingly, from human-induced pluripotent stem cells (hiPSCs). These cell therapies have shown the ability to engraft more successfully and improve electromechanical function of the heart in preclinical studies, including in non-human primates. Advances in cell culture and delivery techniques promise to further improve the engraftment and integration of hiPSC-derived CMs (hiPSC-CMs), while the use of metabolic selection to eliminate undifferentiated cells will help minimize the risk of teratomas. Clinical trials of allogeneic hiPSC-CMs in HF are now ongoing, providing hope for vast numbers of patients with few other options available.


Introduction
Heart failure (HF) represents a significant burden to patients and healthcare systems. It is estimated that HF affects approximately 60 million people worldwide [1] and is the most common cause of hospitalization in the elderly [2]. Patients with HF experience poor quality of life (QoL) [3] and have 5-year survival rates <60%, worse than many common cancers [4,5]. Moreover, the burden of HF is increasing as the population ages, as risk factors such as diabetes and obesity increase in prevalence, and as more individuals survive coronary events, such as acute myocardial infarction (AMI) [6].
Currently available guideline-directed medical and device therapies can only act on and support residual cardiomyocytes (CMs), and prognosis remains poor for many patients.
Although heart transplantation may be an option for patients with advanced HF, this approach is rarely used, partly limited by a shortage of donor organs [8]. Consequently, the ability to generate new CMs and repair the damaged myocardium represents an attractive prospect for helping to improve the prognosis of patients with HF. Cell-based therapies promise to provide patients with new fully functional CMs to repair and/or replace injured heart tissue in patients whose therapeutic options are otherwise limited.
Here, we review the progress to date in the development of cell-based therapies for HF, summarizing early clinical data from double-blind trials of first-generation multipotent cell therapies, before focusing on preclinical data and ongoing clinical trials for next-generation therapies based on human pluripotent stem cells (hPSCs).

A Brief History of Regenerative Medicine for HF
Early cell therapies aimed at treating HF have been based on multipotent cells, which are cells from tissues, such as bone marrow (BM), adult adipose tissue, or the umbilical cord (UC), that can differentiate into multiple cell types within a restricted number of lineages.
A multitude of preclinical studies have assessed the cardiac repair potential of multipotent cells in small and large animal models of myocardial injury. Many of these studies have shown that transplantation of multipotent cells could improve cardiac function [9][10][11]. However, engraftment rates were consistently low, with most transplanted cells quickly lost into the peripheral circulation and the cardiac benefits being moderate or transient [10,11]. The benefits of these therapies would appear to be mediated not by directly replacing the damaged myocardium, but through non-contractile, paracrine effects that help support the function of existing CMs through the release of exosomes, growth factors, and matrix metalloproteinases into the local environment, promoting angiogenesis, and reducing inflammation and fibrosis [10,11] (Figure 1). The modest benefits of these first-generation cell therapies on cardiac function in preclinical models were sufficient to encourage their assessment in double-blind clinical trials in patients following AMI or with ischemic cardiomyopathy or HF (Table 1).   Figure 1. Mechanism of action of next-generation cell therapies versus first-generation cell therapies. In contrast to first-generation therapies, which are largely limited to paracrine effects, nextgeneration therapies aim to promote remuscularization of the heart. BM-MNC, bone marrow-derived mononuclear cell; CM, cardiomyocyte; FGF, fibroblast growth factor; HGF, hepatocyte growth factor; hiPSC-CM, human-induced pluripotent stem cell-derived cardiomyocyte; LV, left ventricle; LVEF, left ventricular ejection fraction; miRNA, micro ribonucleic acid; MSC, mesenchymal stem cell; SDF-1, stromal cell-derived factor-1; VEGF, vascular endothelial growth factor.

Unfractionated BM-Derived Mononuclear Cells
The BM is a source of a variety of multipotent precursors, including mononuclear cells (MNCs), hematopoietic stem cells, endothelial progenitor cells, and mesenchymal stem cells (MSCs). BM-derived MNCs (BM-MNCs) are relatively easy to harvest via BM biopsy/aspiration, and subsequently isolate via density gradient.
BM-MNCs can be collected from BM cells a few hours before administration, without the need to expand or culture cells. Although BM-MNCs can be collected for allogeneic use, the majority of clinical trials have minimized the risk of rejection and the need for immunosuppressants through the use of autologous BM-MNCs (Table 1).

Mesenchymal Stem Cells
MSCs are a subset of heterogeneous non-hematopoietic adult stem cells that express surface markers CD105, CD73, and CD90. Although they originate in the mesoderm, MSCs can self-renew and differentiate into cells of other lineages and not just those from the mesoderm. MSCs can be harvested from various tissues, including BM, adult adipose tissue, and UC, and they are relatively easy to isolate and then expand in vitro, although the cells will eventually senesce in culture. As with BM-MNCs, the transplantation of BM-derived MSCs (BM-MSCs) can be allogeneic or autologous.
BM-MSCs have been extensively studied in double-blind clinical trials in patients with AMI, HF, or ischemic cardiomyopathy. Most studies did not detect significant improvements in LVEF or LV volumes (TAC-HFT [22]; CONCERT_CCRTN [43]) or meet their composite primary endpoints (CHART-1 [38]; DREAM-HF [40,41]). However, there were some indications of potential benefit with this approach in these and other trials. The MSC-HF trial in 60 patients with ischemic HF met its primary endpoint and showed a dose-response relationship with improvements in LVEF and LV end-systolic volume (LVESV), as well as improvements in QoL [34]. Improvements in QoL were also observed in the CONCERT_CCRTN trial [43]. Moreover, at the 4-year follow-up, the BM-MSCtreated patients experienced significantly fewer hospitalizations for angina [35]. The TRI-DENT study assessed the effect of high doses of BM-MSCs (100 million) versus low doses (20 million) in 30 patients with ischemic cardiomyopathy, and it was noted that high doses improved LV function and New York Heart Association (NYHA) class versus lower doses [37]. Although the primary composite endpoint (all-cause mortality, worsening HF, Minnesota Living with Heart Failure Questionnaire [MLHFQ] score, 6-minute walk distance, LVESV, and LVEF) at 39 weeks did not improve in the CHART-1 trial [38], a reduced risk of death or cardiovascular hospitalization was observed with longer-term follow-up in patients with LV end-diastolic volume of 200-370 mL [39]. Similarly, although BM-MSCs did not reduce the risk of the primary endpoint of time to recurrent non-fatal decompensated HF-related major adverse cardiovascular events (HF-MACE) in DREAM-HF, reductions in the risk of other clinical outcomes, such as myocardial infarction (MI) or stroke, were noted [41].

UC-Derived MSCs
UC-derived MSCs (UC-MSCs) offer advantages over BM-MSCs in that they are widely available, and do not require an invasive procedure to harvest. Moreover, they have low immunogenicity [60], and a higher proliferative capacity [61] than BM-MSCs. Clinical trials with UC-MSCs are limited in number, but are predominantly positive.
In RIMECARD, a randomized, double-blind trial of 30 patients with HFrEF, an intravenous (i.v.) infusion of allogeneic UC-MSCs (1 × 10 6 cells/kg) was compared with placebo and shown to improve LVEF, NYHA functional class, and QoL (MLHFQ) [45]. In a trial studying the safety and efficacy of an intracoronary infusion of UC-MSCs (6 × 10 6 cells/kg) in 116 patients with AMI, cell therapy was also shown to improve LVEF, myocardial viability, and decrease in LVESV and LVEDV compared with placebo at 18 months [44]. A double-blind clinical trial in 50 patients with LVEF ≤ 45% who were selected to receive an elective coronary artery bypass graft (CABG) assessed the safety and efficacy of UC-MSCs (1 × 10 8 cells/kg), with or without administration with a bovine collagen hydrogel to aid engraftment and functional integration, with control patients not receiving UC-MSCs. At 12 months, mean infarct size as a percentage of LV mass decreased after treatment with UC-MSCs with collagen hydrogel, but increased with UC-MSCs alone or with no UC-MSCs [46]. This study suggests that supporting the engraftment of cells may provide additional benefits, and the molecular mechanism and retention of UC-MSCs in the heart should be clarified in the future.

Adipose-Derived Regenerative Cells/Adipose-Derived MSCs
Adipose-derived regenerative cells (ADRCs) are a heterogeneous population of multipotent cells, including MSCs, obtained from the vascular stromal fraction of adipose tissue [62]. The adipose-derived MSCs are more abundant than BM-MSCs and harvesting (via liposuction) is arguably less invasive than BM aspiration. Moreover, ADRCs do not require culture or expansion.
Preclinical trials have shown beneficial effects in animal models of ischemic cardiomyopathy [63][64][65], but the results from double-blind clinical trials are limited and mixed. The PRECISE trial examined the safety and feasibility of administering ADRCs in 27 patients with coronary artery disease not amenable to revascularization, and ADRC treatment, but not placebo treatment, was associated with a significant increase (p < 0.001) in LV total mass from baseline to 6 months. In addition, LV infarcted mass increased with placebo (p = 0.01) but not ADRC treatment. However, there were no significant changes in LVEF or LV volume with either treatment. The PRECISE trial was limited by a small sample size and imbalances in baseline magnetic resonance imaging (MRI), single-photon emission computed tomography (SPECT) measurements, and age between treatment groups [47].
The ATHENA I and II trials aimed to assess the effects of ADRCs in patients with chronic ischemic cardiomyopathy (LVEF ≥ 20% to ≤ 45%). Enrollment in these trials was terminated prematurely due to cerebrovascular events deemed unrelated to the cell product. In those patients who were enrolled, no improvements in LV function or volume were observed with ADRCs; however, an improvement in QoL (MLHFQ) was reported with ADRCs [48].
Two completed Phase 2 trials of allogeneic ARDCs have yet to publish their results: SCIENCE and CSCC_ASCII [56,57].

C-Kit-Positive Cardiac Cells
C-kit-positive cardiac cells (CPCs) are multipotent, clonogenic stem cells with subpopulations that can preferentially differentiate into myocytes or endothelial cells. Treatment with these cells has been shown to promote cardiac regeneration and angiogenesis via paracrine effects in animal models [66,67].
Several clinical trials have assessed the potential effects of CPCs in patients with HF. The CONCERT_CCRTN trial assessed the safety and efficacy of autologous CPCs, BM-MSCs, and a combination of BM-MSCs and CPCs, versus placebo in 125 patients with ischemic HF [43]. Interestingly, CPCs were noted to reduce HF-MACE over 12 months; however, improvements in LV function and reductions in scar size were not noted. Thus, the mechanism for the reduction in HF-MACE in the CONCERT_CCRTN trial is unclear. SCIPIO, a Phase 1 trial assessing the effect of CPCs in patients with post-infarction LV dysfunction before CABG, reported encouraging efficacy results, with increases in LVEF and decreases in scar size [68]. However, the publication was later retracted due to doubts over the reliability of the work performed by the laboratory that had prepared the cells [69].

Summary of First-Generation Cell Therapies
Preclinical studies and double-blind clinical trials have confirmed the safety of many first-generation stem cell therapies in patients with HF or ischemic cardiomyopathy. Firstgeneration cell therapies, such as BM-MNCs, appear to have low engraftment rates, and functional benefits appear limited and mediated largely by paracrine effects supporting existing CMs, rather than an ability to integrate and regenerate new myocardium. Approaches that aid stem cell engraftment and promote functional integration, may therefore be required to observe consistent clinical benefits with these categories of cell therapies in patients with HF.

Next-Generation Stem Cell Therapies
For regenerative therapies to realize their potential in HF, the new cells must not only have paracrine effects, but also survive, engraft, create gap junctions, and couple electrically with native CMs. By using pluripotent stem cells (PSCs)-which can differentiate into all cell types-it is possible to culture LV-specific CMs [70]. PSC-derived cell products represent a new generation of cell therapies based on the transplantation of mature cell types that may be more likely to engraft and integrate electrically with the host myocardium compared with first-generation multipotent stem cell therapies.
There are two types of PSCs: embryonic stem cells (ESCs) and induced PSCs (iPSCs).
ESCs are derived from the inner cell mass of blastocysts and can differentiate into all three embryonic germ layers. iPSCs are generated from fully differentiated adult somatic cells, e.g., skin fibroblasts or peripheral blood cells. Somatic cells are reprogrammed to become PSCs, usually by overexpressing the transcription factors required for pluripotency [71,72].

Human PSC-Derived CMs in Preclinical Models of HF
Studies of rodent models of myocardial injury have reported the beneficial effects of human ESC (hESC)-derived CMs (hESC-CMs). Injection of 1 × 10 6 hESC-CMs into the hearts of immunocompromised mice following an MI induced by ligation of the left anterior descending (LAD) coronary artery, resulted in improvements in LV function at 4 weeks, although not at 12 weeks [73]. This short-term benefit is likely reflective of paracrine effects. Although hESC-CMs were shown to integrate and mature in vivo, it was suggested that graft size may have limited the longer-term functional benefit in this study. Another study has noted a longer-term benefit with an intramyocardial injection of 1 × 10 6 hESC-CMs in mice after induction of an MI: LVEF was improved at Day 28 and Day 60, scar size and CM apoptosis were significantly reduced, and CM proliferation, capillary bed, and arteriole number all increased [74]. hESC-CMs transplanted into a guinea pig model of cardiac injury have also been reported to improve the mechanical function of the heart and reduce ventricular tachycardia [75]. Moreover, grafts were heterogeneous, with uncoupled regions and regions that contracted synchronously with the host heart [75], suggesting a level of electromechanical integration, but also showing a need to further optimize engraftment. Similar to studies with hESC-CMs, human iPSC (hiPSC)-derived CMs (hiPSC-CMs) have also shown some benefits in rodent models. An intramyocardial injection of 10 × 10 6 hiPSC-CMs into the myocardium 10 days after ligation of the LAD coronary artery, resulted in reduced mortality and cardiac remodeling versus controls, and LVEF increased by almost 20% after 4 weeks [76]. Grafted CMs could also be detected 1 month after transplantation in this study. In a rat model of HF, a tissue-engineered patch embedded with hiPSC-CMs and human neonatal fibroblasts was grafted onto the epicardial surface covering the infarcted tissue, and electrical activity was found to be improved and end-diastolic pressure reduced after 3 weeks [77].

Human Pluripotent Stem Cell Cardiomyocytes in Large Animal Models of HF
It is notable, however, that rodent hearts show marked differences in anatomy and physiology compared with human hearts, such as a much faster heart rate. Consequently, cell-based therapies should also be tested in large animal models to provide a better indication of efficacy and safety. In a porcine model of AMI, intramyocardial injections of three cell types derived from hiPSCs (CMs, endothelial cells, and smooth muscle cells) were administered through an epicardial fibrin patch loaded with insulin growth factor 1 to promote survival. This approach was shown to result in engraftment and improved LV function after 4 weeks, without inducing ventricular arrhythmias [78]. Another porcine model of MI has demonstrated stable engraftment that formed vascular networks and resulted in a large degree of remuscularization in the heart after transplantation with hESC-CMs. Although no teratomas were observed in that study, ventricular tachyarrhythmias were observed [79]. Studies in non-human primates have also produced promising findings. When 1 billion hESC-CMs were injected into the myocardium of immunosuppressed macaques 2 weeks after induction of an MI, significant remuscularization of the infarcted myocardium was noted. Grafts were shown to have developed electromechanical junctions and showed synchronization of calcium transients to the electrocardiogram from the host myocardium, indicating electromechanical coupling. In contrast to small animal models, however, non-fatal ventricular arrhythmias were also observed [80]. Another study of 750 × 10 6 hESC-CMs transplanted into a macaque monkey ischemia-reperfusion model of MI has demonstrated improvements in LVEF at 1 month and 3 months post-transplantation. Grafts were shown to have formed electromechanical junctions with the host myocardium, but a subset of animals were also noted to experience graft-associated ventricular arrhythmias [81]. In immunosuppressed cynomolgus monkeys, an intramyocardial injection of 4 × 10 8 allogeneic iPSC-derived CMs (iPSC-CMs) 14 days after a 3-hour occlusion of the LAD coronary artery, resulted in improved contractile function at 4 and 12 weeks. Moreover, the grafts survived for 12 weeks and showed electrical coupling with the host CMs. This study reported an increased incidence of ventricular tachycardia with iPSC-CM treatment compared with vehicle-treated controls, but this was transient [82].
Preclinical studies in animal models have shown that hPSC-derived CMs (hPSC-CMs) can engraft into the host myocardium, grafts can be sustained over several months, and can achieve electromechanical coupling with the host myocardium and improve LV function. Although some of the benefits of hPSC-CMs may be due to paracrine mechanisms, the presence of myocardium remuscularization and electromechanical coupling indicate the potential for benefits due to direct interactions between hPSC-CMs and host CMs.

Challenges for hPSC-Based Regenerative Therapies in HF
For next-generation hPSC-based regenerative stem cell therapies to be tested and used in patients with HF, several concerns and challenges need to be addressed, including the potential risk of teratomas and arrhythmias, the need for an optimal delivery system and improved engraftment rates and survival, as well as large-scale production.

Teratoma Prevention
Teratomas are tumors made up of tissues from multiple germ layers. The ability of undifferentiated PSCs to form any cell type means that they form teratomas after transplantation [83,84]. Many preclinical studies have not observed teratoma formation following the administration of hPSC-CMs [75,76,80,81]. However, the true incidence of teratomas may be under-represented in some preclinical studies, which have often used relatively few animals and relatively short follow-up. Moreover, even a small risk may be clinically significant if millions of cells are injected. Indeed, undifferentiated hPSCs can give rise to teratomas even if only 0.025% of residual undifferentiated hPSCs remain [85]. There is therefore a need to develop technologies to aid early detection of teratomas, and it has been suggested that a combination of biomarkers (α-fetoprotein, carcinoembryonic antigen, and human chorionic gonadotrophin) along with an MRI, may provide a sensitive approach for identifying teratomas from hPSCs [86].
In addition to improved detection of teratomas with hPSCs, it is important to prevent teratoma formation through optimized pre-implantation protocols. One approach to limit the potential for teratomas with hPSCs, is to purify cultures to remove any undifferentiated cells before administration. Multiple approaches have been assessed to help identify and remove undifferentiated cells ( Table 2). Use of a monoclonal antibody against cell surface antigens specific to hPSCs, can allow separation of cells through fluorescence-activated cell sorting [87]. However, cell sorting may be impractical when large numbers of cells are required. The use of small molecule inhibitors may also reduce the risk of teratomas, inducing the selective apoptosis of undifferentiated hPSCs [88]. For example, survivin is an anti-apoptotic factor specific to hPSCs, and chemical inhibitors of this factor, such as quercetin or YM155, have been reported to promote cell death in undifferentiated hPSCs, but not differentiated cells [89]. Treatment of in vitro cultures with brentuximab vedotin, which targets CD30-positive hiPSCs, has also been reported to promote cell death of nondifferentiated hiPSCs and reduce teratoma formation in mice [90]. Another approach to eliminating undifferentiated hPSCs, is through metabolic selection ( Table 2). Fatty acid synthesis is important for the survival of undifferentiated hiPSCs, but not hiPSC-CMs; consequently, inhibition of cells with fatty acid synthase before transplantation represents an approach for eliminating undifferentiated cells and minimizing the risk of teratomas [91]. Undifferentiated hPSCs use glutamine and glucose to produce energy, but cannot use lactate [92]. In contrast, hPSC-CMs can use lactate as an energy source. By culturing cells in a glucose-and glutamine-free medium supplemented with lactate, undifferentiated hPSCs can be eliminated to the level of <0.001% [88,92,93]. Glucose can inhibit maturation of hPSC-CMs [94], and therefore metabolic selection by restricting glucose may also aid the maturation of CMs during purification. Methionine is also required in large amounts by hPSCs, and prolonged depletion of methionine can lead to selective apoptosis of hPSCs [95]. Metabolic selection of hPSC-CMs from undifferentiated cells represents an approach that can be used on large-scale cultures and with limited requirements for specific or expensive compounds [88]. Lectin-toxin fusion protein [103] Binds to hPSCs only and delivers cytotoxic protein when internalized, eliminating hPSCs Clostridium perfringens enterotoxin [104] Toxic that binds to Claudin-6, a tight-junction protein specific to hPSCs, and kills undifferentiated cells

Risk Reduction of Arrhythmia after Transplantation
Electrical integration of the grafted cells into the myocardium is an important goal of regenerative cell therapy, and engraftment arrhythmias represent an obstacle to their use clinically. Previous studies have reported the development of ventricular arrhythmia after transplantation of hPSC-CMs into the hearts of larger animal and non-human primate models [79,80,82,107]. These arrhythmias typically occur within the first two to three weeks after transplantation of hPSC-CMs and then may persist/reappear for up to a month, after which the heightened risk for new events disappears [79]. A study using electrical mapping and pacing suggested that the mechanism of ventricular tachycardia after transplantation is automaticity rather than macro-reentry. Contamination of atrial cells, pacemaker cells, and non-ventricular CMs may cause arrhythmias [79]. Cell dose, injection volume, cell condition, and cell retention rate may also be important.
There are various potential strategies for the prevention of arrhythmias. Ensuring that transplanted cells are purified and do not include non-CM cells may aid electromechanical coupling. Moreover, transplanting hiPSC-CMs of a ventricular phenotype with electrophysiological characteristics close to those of the host tissue may also aid electrical integration. The initial hPSC-CMs were electrophysiologically immature. For example, the resting membrane potential is less hyperpolarized in immature hPSC-CMs (approximately −60 mV, similar to that of nodal cells) than in mature ventricular CMs (approximately −90 mV). Furthermore, immature hPSC-CMs express high levels of hyperpolarizationactivated cyclic nucleotide-gated channel 4 in the plasma membrane, which is characteristic of pacemaker cells. These aspects make it easier for immature hPSC-CMs to beat spontaneously (i.e., to show automaticity), whereas adult ventricular CMs are electrically quiescent until triggered by the depolarization of adjacent cells [108]. Therefore, transplantation of more mature ventricular CMs may be useful for the reduction of arrhythmogenic risk. Studies have shown that differentiation and purification protocols can produce CMs of the ventricular phenotype for transplantation [109]. Ensuring a high survival rate of grafts is also important as necrotic tissue may cause inflammation and serve as a substrate for arrhythmias.
When large numbers of hiPSC-CMs were transplanted, single floating frozen cells were commonly used. These cells were thawed just before usage, but the cell surface proteins (ion channels, growth factor receptors, cell adhesion molecules, etc.) of these cells were destroyed by enzyme digestion, and the freeze-thaw process impaired cell survival after transplantation. As a result, the cell engraftment rate became extremely low, and the dead cells had the potential to cause local inflammation and injury to the surviving CMs or host CMs, resulting in induction of automaticity arrhythmia. Transplantation of a large volume of hiPSC-CMs at one site is not desirable, as it may destroy the physiological electrical conduction system of the host CMs. Finally, myocardial damage from intramyocardial injections could also trigger arrhythmias. Intramyocardial transplant injection devices that efficiently and safely introduce and distribute hiPSC-CM aggregates/spheroids, which have been reported to improve cell survival, engraftment, and cardiac function in rodents and pigs versus suspensions of single cells [99,107] are also in development [110]. We deduced that the improvement in transplantation techniques may greatly reduce graftinduced arrhythmia. Another approach to minimize the impact of engraftment arrhythmia is to employ pharmacologic approaches. Indeed, ivabradine and amiodarone have been used to effectively suppress engraftment arrhythmia in a porcine model of MI treated with hPSC-CMs [111].

Optimizing Delivery
There are numerous routes for administering hiPSC-CMs (Figure 2). The standard procedure for introducing cells into the heart is to inject them via intramyocardial (usually transendocardial or transepicardial) or intracoronary routes. Intracoronary delivery has the advantage of being less invasive than approaches requiring surgery, such as the placement of patches or transepicardial injections. Intracoronary injection may be unsuitable for delivering larger cells, such as MSCs, which could occlude the microcirculation and for use in patients with HF who have highly diseased arteries. A study in pigs has suggested that retention of peripheral blood MNCs is better after intramyocardial injection (11 ± 3%) than after intracoronary injection (2.6 ± 0.3%), with a smaller proportion of cells leaving the heart and entering the pulmonary circulation (intramyocardial injection, 26 ± 3%; intracoronary injection, 47 ± 1%) [112]. Indeed, it has been noted that 1 h after intracoronary injection, only 2-5% of cells were detected in the heart, with the majority found in the liver and spleen [113]. Graft survival is also poor following intracoronary administration of dispersed hPSC-CMs [114]. Although intracoronary injection of hPSC-CM aggregates can lead to partial engraftment, cardiac ischemia can develop and result in scars similar in size to the injected spheroids [114]. Another challenge for intracoronary delivery, is that hPSC-CMs would need to migrate from the vasculature into the myocardium.
Intramyocardial injections offer several other advantages over intracoronary administration, such as the ability to target cells to the myocardium and to a specific location, and the delivery of larger cells or aggregates/cardiospheres that might otherwise occlude microvessels. However, specialist training may be required for intramyocardial injection, and there is potential for perforation and myocardial damage [115]. Most clinical studies have used intramyocardial delivery, usually in the form of transendocardial catheter injections [115].

Further Improvement in Engraftment Rates and Longevity
Engraftment rates with hPSC-CMs still remain relatively low (e.g., no grafted hESC-CMs could be detected 4 months after administration of a fibrin patch loaded with hESC-CMs in a rat model of HF [116] or 140 days after administration of a cell suspension in a cynomolgus monkey model of MI [117]). Therefore, there is a need to further improve engraftment. It has been suggested that engraftment could be improved through tissue engineering and alternative methods of transplantation [118].

Cardiospheres
Suspensions of single stem cell-derived CMs tend to graft poorly. The formation of PSC-derived CM (PSC-CM) aggregates/spheroids through cell-cell adhesion has been reported to improve cell survival when injected into mouse hearts [99]. Intramyocardial injection of spheroids-made up of approximately 1000 hPSC-CMs-into the infarcted hearts of rodents and pigs, produced significantly better engraftment and greater improvements in cardiac function versus suspensions of single cells [107]. PSC-CM aggregates/spheroids were generated in a floating cell condition, which means that they do not require enzyme digestion for harvesting the cells; cell surface proteins (such as ion channels, growth factor receptors, cell adhesion molecules), as well as extracellular matrix and matrix-bound growth factors are intact, which in turn greatly improves cell retention after transplantation. Conventional needles have a beveled edge at the tip, to cut the tissues and microvessels at the injection site, resulting in bleeding and spheroid leakage. Injection of spheroids into the myocardium of pigs by a specially designed needle with a cone-shaped tip and multiple side holes (SEEDPLANTER ® ) resulted in reduced tissue damage and bleeding, and better retention of spheroids within the myocardium than use of a conventional needle [110].
Culturing hPSC-CMs as spheroids may also lead them to acquire a more mature phenotype, which could improve engrafting and electrical coupling with native CMs. Coculturing hiPSC-CMs with endothelial cells, smooth muscle cells, and cardiac fibroblasts in a three-dimensional (3D) environment, yielded spheroids that contained all four cell types, and hiPSC-CMs have a more adult-like phenotype than those produced in two-dimensional (2D) cultures [119].

Delivering Cells via Epicardial Patches/Sheets
An alternative approach for the transplant of iPSCs is to use tissue engineering to produce sheets of cells, or 'patches', with specific architecture mimicking the structure that biological tissues achieve via encapsulation of cells in an extracellular matrix. These patches can be attached to the epicardial surface of the heart with adhesives or sutures [118,120].
Patches may be developed on a scaffold of natural or synthetic materials. Patches of hiPSC-CMs developed on a fibrin scaffold have been shown to improve engraftment and LV function compared with a suspension of single cells when transplanted onto the ventricle in a porcine model of MI [121]. A scaffold of polylactic-co-glycolic acid, a synthetic polymeric material, has also been used to develop a patch of iPSC-CMs on a large scale, and this approach has been reported to improve LVEF in a porcine model of ischemic cardiomyopathy [122]. Mesh-structured engineered heart tissue patches made up of iPSC-CMs have also been reported to improve LV function and establish dose-dependent remuscularization of guinea pig hearts [123].
Delivery of iPSC-CMs via patches offers some advantages, in that surgeons can visually confirm attachment and positioning. Attachment of the patch may cause less damage than an intramyocardial injection, and the patch provides a structural environment that may promote engrafting. Patches also have potential disadvantages in that their application may be more invasive than catheter-based delivery. Moreover, the epicardium and pericardial adipose tissue on the ventricular free wall [118,124] may present barriers that interfere with the full integration of the cells into the host myocardium. Epicardial patches may also be separated from the host myocardium by scar tissue, which may hinder electrical coupling with host CMs. Although several studies have noted functional and electrical recovery after grafting of iPSC-CM cell sheets/patches [121,122,125], it was reported that hESC-cardiac tissue patches introduced into a rat model of HF were electromechanically active, but were not electrically coupled to the host CMs at 4 weeks. In contrast, cells introduced via intramyocardial injection were electrically coupled to the host [120].
Further options for improving engraftment may be to utilize a combination of approaches. The injection of hPSC-CMs into the myocardium, accompanied by placement of an MSC-loaded patch on the epicardium has been noted to improve cardiac repair in rats [126]. The MSC patch released paracrine factors that enhanced vascular regeneration, and also significantly improved the retention and engraftment of intramyocardial injected hiPSC-CMs.

Economic Improvement of Production
It is estimated that approximately several hundred million to one billion CMs would be needed to completely replace the CMs lost in the LV of a patient with severe HFrEF [88]. Therefore, the production of hiPSC-CMs needs to be scalable to meet the demand to conduct trials and to treat patients if shown to be effective. Currently, initial culturing of hiPSCs and hiPSC-CMs can be performed efficiently and on a large scale using a 2D culture system [109], with cardiosphere development occurring in microwell plates after differentiation and purification [107] (Figure 3). It has been suggested, however, that 3D culture techniques may offer greater scalability, producing larger numbers of cells than traditional 2D cultures [88,127]. Moreover, 3D cultures allow iPSC-CMs to develop a more mature phenotype than 2D monolayers [119], possibly due to the low oxygen environment [128]. 3D suspension cultures may also be more economical, as there is no requirement to use expensive cell-adhesive coating proteins. Although massive 3D suspension culture systems offer the production of great numbers of hiPSC-CMs, there is a need to confirm the quality of hiPSC-CMs manufactured in the process, particularly in terms of the purity of CMs to minimize the risk of teratoma formation [129]. Metabolic purification systems that restrict glucose and glutamine and supplement lactate offer an approach that may allow the purification of hiPSC-CMs in massive 3D suspension culture systems.

Clinical Trials with hPSC-CMs
One small trial to assess the safety and feasibility of using hESC-derived cardiac progenitor cells (CPCs) to treat HF has already been completed: the ESCORT trial [70]. This trial assessed the efficacy of a fibrin patch embedded with hESC-derived CPCs implanted on the epicardium during CABG. Six patients with LVEF ≤ 35% and a history of MI were treated. No patients showed arrhythmias or developed teratomas during follow-up, but three patients showed clinically silent alloimmunization. At the 1-year follow-up, all patients assessed showed a reduction in HF symptoms. A significant increase in heart wall motion was also seen in cell-treated areas, along with a non-statistically significant increase in LVEF.
Early phase trials to confirm the safety and efficacy of hPSC-CMs in HF are now ongoing (Table 3). These trials are relatively small (10-55 patients), with most being open-label and very few having a control arm. The trials are predominantly assessing the effect of hPSC-CMs in patients with ischemic HFrEF, although two studies also include patients with non-ischemic HFrEF. The number of transplanted cells varies, probably due to differences in cell purity, cell transplantation form, and engraftment rate. The primary objective of most of the trials is to assess safety. Assessment of LVEF or wall thickness by echocardiography are the primary objectives in only two studies; however, most other studies include echocardiography and MRI assessments of efficacy as secondary endpoints. Moreover, several studies have also included functional (6-minute walk distance/time) and QoL (MLHFQ) assessments.  Although preclinical trials have used both hESCs and hiPSCs, the current regulatory environment and potential ethical issues related to the use of hESCs means that the focus of most trials is on the use of hiPSCs (Table 3). Ongoing trials are also using allogeneic rather than autologous cells. There are several reasons why allogeneic cells may be preferred over autologous cells. The function of cells for autologous use in patients with HF may be compromised by age or comorbidities, or genetic disorders in the cases of some hypertrophic or dilated cardiomyopathies. In addition, allogeneic cells do not require harvesting, reprogramming, or quality checking for each host, and therefore, their production can occur more rapidly and on a larger scale than autologous cells. Autologous cells have some advantages over allogeneic cells in terms of improved engraftment and reduced risk of rejection, and the lack of requirements for immunosuppressants. The use of autologous cell therapy would thus be beneficial for patients with HF who are not tolerant of immunosuppressants. However, a more rapid and efficient process for obtaining, differentiating, and checking hiPSC-CMs from each patient would need to be established first.

Summary
First-generation cell-based therapies using multipotent cells demonstrated safety in preclinical models of HF, but poor engraftment rates and a limited ability to couple electrically with existing CMs meant that improvements in cardiac function in clinical trials were largely limited to those attributable to paracrine effects. Next-generation stem cell therapies using CMs derived from hESCs or, increasingly, from iPSCs, are in development and have shown the ability to engraft more successfully, and to improve electromechanical function of the heart in preclinical studies, including in non-human primates. These nextgeneration therapies are being enhanced by advances in techniques to improve engraftment rate and to minimize the risk of teratomas by purifying cells on a large scale. Clinical trials of allogeneic hiPSC-CMs in HF are now ongoing, providing hope for vast numbers of patients with few other options available.
Author Contributions: Writing-original draft preparation, Y.K.; writing-review and editing, K.F. All authors have read and agreed to the published version of the manuscript.
Funding: This review and the article processing charges were funded by Novo Nordisk A/S.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.