Heart failure (HF) is a disease in which the ventricular filling and/or myocardial contraction are compromised due to irreversible pathological modelling in the cardiac structure and function (1–3). Ischemic heart disease is the most common cause of HF. More than 26 million people are living with a failing heart worldwide (4), and 1 in every 8 reported deaths was due to HF (5). This prevalence is expected to increase by 46% in 2030 (6, 7). The most common cause of HF is ischemic heart disease. Myocardial ischemia as a result of the blockage in the coronary artery can affect myocardial contractility, and electrical conduction, and alters cardiac energetics (8). Pathological hypertrophic remodelling of the left ventricle, fatal cardiac tachyarrhythmia (9) followed by a cascade of secondary damage including inflammatory reaction, myocardial cell rupture and fibrotic scarring could collectively cause an abrupt decrease in left ventricular ejection fraction (LVEF) and failure. Despite technological advancement and all the current pharmacological-based treatment options, effective therapy that could prevent damaged hearts from remodelling and failure is still lacking. Heart transplantation is the only cure for end-stage heart failure, but the therapy is challenged by the lack of donor and graft rejection.

Stem cells have been the research interest and hope to remuscularize the weakening, injured myocardium and reverse cardiac remodelling. Compelling evidence has shown that pluripotent stem cells (hPSCs), such as embryonic stem cells (ESCs) or induced pluripotent stem cells (hiPSCs) offer the indefinite source of cardiomyocytes and are the only cells which can be scaled up to produce a clinical relevant number for remuscularizing injured myocardium. Studies have shown that intramyocardially injected hPSCs-cardiomyocytes engrafted, integrated synchronously with the host myocardium, regenerated the remodelled, thin myocardium and improved cardiac function. Whilst the clinical benefits of other adult stem cells such as mesenchymal stromal cells or cardiac-derived progenitor cells in cardiac regeneration are widely acknowledged (10, 11), the extent of remuscularization in the injured hearts was generally far more encouraging using hPSC-cardiomyocytes, as demonstrated in the past studies (12–15).

To establish a suitable disease model for reliable experimentation, investigators need to take into account the animal species of choice, the subject availability, the cost and the similarity/difference between the human subject as well as the methods to induce myocardial injury that produces the appropriate pathological microenvironment mimicking clinical conditions which is suitable for testing therapeutic remuscularization intervention. Table 1 summarizes the different considerations in the use of small animals (mice, rats) and large animals (pigs, dogs, monkeys) for modelling ischemic heart disease.

An ideal disease model should mimic the pathophysiology in humans in order to more accurately and reproducibly examine any novel therapy that would successfully translate into clinical applications. In preclinical research, any new therapy would first be tested in small animal models. This is because small animals such as rodents have a short life cycle, hence requiring low maintenance cost and high availability are instrumental in producing statistically meaningful analysis within a comparatively shorter time over the use of large animals (Table 1). An early rat HF model was established by Pfeffer et al. using coronary artery ligation (16). The groundbreaking study dated back to 1979 also served as an important foundation for the development of the successfully translated drug Captopril, the angiotensin-converting enzyme that profoundly improved heart function and survival in post-myocardial infarction (MI) patients (17).

Nonetheless, the differences in heart anatomy, size, hemodynamic characteristics and responses to drugs or treatment between small animal models and humans are likely the cause of failed translation to clinical trials. These differences could also explain the rather disappointing clinical results observed in most major human stem cell therapy trials, despite the overwhelmingly positive outcome and optimism reported in laboratory small animal studies. In 2016, Zwetsloot et al. presented a systematic review and meta-analysis of preclinical studies involving cardiac stem cell treatment in MI animal models (18). They concluded that the magnitude of effects of CSC treatment in the small animal MI model was found to be greater than that of large animals. Moreover, the recent incidence of cardiac arrhythmias reported from non-human primate and pig studies following human cardiomyocyte transplantation was not previously observed in small animals possibly due to their high heart rate (13, 19, 20). Hence, small animal studies may offer important preliminary insights about the tested treatment, but a reassessment of its therapeutic efficacy must be performed on large animals whose systems are physiologically and anatomically more resembles humans prior to starting clinical trials. This is also in line with the recommendations by the transnational alliance for regenerative therapies in cardiovascular syndromes (TACTICS) international group that large mammals should be used as confirmatory studies in view of their resemblance to human disease (21).

Many strategies have been introduced to create MI and HF models in large animals, with the primary objective to occlude major coronary vessels and induce ischemic injury. These include invasive thoracotomy-enabled permanent left anterior descending (LAD) coronary ligation, reversible LAD coronary artery ligation ischemic reperfusion-induced myocardial injury, coronary micro-embolism, hydraulic occluder or ameriod constrictors, or less-invasive percutaneous transluminal coronary angioplasty (PTCA) balloon occlusion-reperfusion of the coronary artery (Table 2). Nevertheless, the most common strategies employed in recent preclinical remuscularization studies using large animals were mainly the thoracotomy/permanent LAD ligation (24) or ischemic reperfusion (30) and PTCA-assisted balloon occlusion/reperfusion of LAD contrary artery (20). Unlike others, these methods allow occlusion to take place at the specific location of the LAD coronary arteries and produce predictable, consistent infarct size which is pivotal for remuscularization study. Notably, the mortality of the LAD occlusion method in large animals is considerably high as they are prone to surgical-induced trauma, high risk of bleeding and developing fatal ventricular fibrillation following MI (35).

For assessing the efficiency and efficacy of cardiomyocyte therapy on cardiac remuscularization, the animal model must develop clear infarction to create the need for cellular reconstitution. Methods such as coronary microembolism, pacing-induced tachycardia and toxic injury can induce dilated cardiomyopathy even without the presence of clear infarcts, making them less common methods used for the cardiac remuscularization-related study (36).

The surgical procedures required to perform in vivo cardiac remuscularization studies are mostly invasive, some of which require a thoracotomy, percutaneous coronary intervention, and for enabling direct intramyocardial cell injection. Handling large animals to prepare them for these procedures is challenging and inhumane without proper and effective perioperative sedation, general anaesthesia and analgesia. The outcome of infarction in swine was also reported to be affected by the choice of anaesthesia and breed (72). Employing a suitable anaesthetic strategy during MI modelling in the large animal is important for achieving adequate anaesthesia during the induction of the intended pathophysiological changes, as well as securing stable post-surgery hemodynamic and recovery, and reducing the risk of anaesthesia-related mortality due to malignant hyperthermia or complications following long anaesthesia (73). Here, the choice of pre-emptive sedatives, anaesthesia and analgesia used in large animal studies in Table 2 are discussed.

Successful remuscularization of an injured heart requires promising regeneration of the dead myocardial tissue in order to restore the myocardial muscle density and contractile strength. Many cell candidates have been tested in laboratories or clinics (10), but this review will only discuss the past animal studies that used human cardiomyocytes derived from hiPSCs. Methods and the efficiency of deriving functional cardiomyocytes from human PSCs have been improved substantially, either by using growth factors, small molecules, or combinations of both at a defined culture time (99).

A myriad of methods has been introduced to administer cells into the heart, with an aim to maximize cell homing, retention, engraftment and subsequent survival and function. Previously, cell administration via the systemic intracoronary route in cynomolgus monkeys has proven inefficient with a high incidence of embolism and poor graft survival (100). Local intramyocardial route, however, is the most favourable method to deliver cardiomyocytes, either via trans-endocardial or trans-epicardial injection. The trans-endocardial route is considerably less invasive than other local injection methods as it can be achieved via the percutaneous catheter. The unique challenge in implanting cells into the myocardium is the characteristic of the heart being a constantly contracting organ, which creates the mechanical force that squeezes the injected cells out through the needle track (the “washout” effect) or the broken blood vessels because of direct injection (101, 102).

To achieve high cell retention, epicardial implantation of a cardiac patch may be considered. The only shortcoming of this delivery method is the inevitable, invasive thoracotomy required for the implantation. Recently, a minimally invasive intrapericardial cell injection was proposed (102). The authors performed the procedure through two small incisions (one for insertion of a camera probe and another for a needle with exosomes in hyaluronic acid hydrogel) on the pig chest wall and showed minimal inflammation. They also tested the delivery method using hiPSC-derived cardiac progenitor cells in decellularized porcine heart matrix hydrogel and demonstrated promising cell engraftment on the epicardial surface, minimal immune response as well as the evidence of in vivo cardiomyocytes differentiation of the injected cardiac progenitors. However, this reported benefit was only tested and observed in rat hearts.

Studies have shown that the number of transplanted cardiomyocytes determines the degree of remuscularization in the injured heart (103). The range of cardiomyocyte numbers which were tested in the∼40 g macaque infarcted hearts was between 400 and 1,000 million human induced pluripotent stem cells-derived cardiomyocytes, through intramyocardial injection (12, 20, 30). Assuming the average monkey body weight in those studies was 9 kg, the dose for every kg body weight in a human would be 44–111 million cells or 3–7 billion cells in an adult human with an average body weight of 70 kg. However, these numbers remain inconclusive, as a question was raised about the clinical relevance in proportion to the size of a human heart of which the left ventricle contains only∼5 billion cardiomyocytes (104).

Another common approach in preclinical CM transplantation studies in addition to intramyocardial injection is by epicardial implantation of engineered heart tissue. A recent study by Querdel et al. (2021) showed that EHT made of high cardiomyocyte dose (1.5 × 2.5 cm, 12 × 10⁶ cardiomyocytes) improved heart function in guinea pigs, with in situ time-dependent cardiomyocyte proliferation within the implanted EHT (103). The authors also claimed to have successfully upscaled to generate a 7 × 5 cm human-relevant-sized EHT with 450 million cells for clinical use. In a MI study using Gottingen minipigs (weight 20–25 kg each), Suzuki et al. (2021) transplanted four large, 2.5 × 2.5 cm cardiac tissue made with 2.5 million cardiomyocytes on the aligned nanofibers to the infarcted myocardium (1 billion cells in total, 50 million kg⁻¹ for a 20 kg minipig) (25). They concluded the treatment improved cardiac function and angiogenesis with antifibrotic effects but low engraftment, possibly due to immune rejection.

In most cases, the established MI animal models used for testing the regenerative capability of any cell candidate were from xenogenic sources, e.g., human cardiomyocytes to swine or macaques' hearts. One of the key determinants of successful clinical use of cardiomyocyte therapy is dictated by the degree of engraftment and survival of the transplanted xenogenic cells and this outcome is affected by the immunologic responses of the host recipient upon transplantation. Most of the preclinical cell therapy experimentation involves xenotransplantation (12). Transplantation of non-autologous cells can result in immune reactions that are primarily caused by acute cellular rejection, mainly because of the T cell alloantigen recognition of the major histocompatibility complex (MHC) (105). Some allogenic cell candidates may have the ability to evade immunorecognition and avoid graft rejection, like the mesenchymal stromal cells (MSCs). These cells are known to be immunomodulatory-privileged and can effectively modulate the immune system by inhibiting T cell proliferation or maturation after transplantation (106). This suggests that MSC transplantation might not need immunosuppression even if the cells are of an allogenic source (107). However, such privilege was found to be withheld when the cells were differentiated (108).

The very original concept of the creation of human iPSCs was the possibility to derive them from autologous sources, and offering cell therapy with the patient's own cells for transplant would resolve the problem with immunorejection (109, 110). However, the high cost underlying each iPSC line generation and the significantly longer time (months) required for differentiation and up-scaling may not be feasible for some clinical conditions which in need of immediate treatment. Hence, getting a universal human iPSC line that could serve “off-the-shelves” would make cell therapy more readily available for the use of broader patients.

The idea of cryobanking human induced pluripotent stem cells (iPSCs) generated from HLA homologous donors matching for human leukocyte antigen (HLA)-A, HLA-B, and HLA-DR alleles has been first advocated in the United Kingdom (111). Instead of using autologous cells, matching the compatibility of the allogeneic iPSCs based on these three most notorious triggers of immune rejection would turn the cell lines transplantable for a larger patient population with good graft survival (111, 112). However, this approach may only apply to countries in which the population has low diversity in HLA haplotypes (113). Moreover, the conflicting finding was also observed in iPSC-derived cardiomyocytes which revealed the need for immunosuppressants despite the reduced immune reactivity due to MHC matching (114).

In 2019, Schrepfer's laboratory revealed that generating allogeneic human iPSCs with hypo-immunogenicity is in fact possible (115). These allogeneic human iPSCs inactivated their major histocompatibility complex MHC class I and class II genes (B2M and CIITA, respectively), as well as upregulated the non-MHC ligand CD47 to silence innate immunity. They tested the hypoimmunogenic human iPSCs in allogenic humanized NSG-SGM3 mice and showed the successful formation of teratoma. The iPSC differentiated derivatives, endothelial and cardiomyocytes from the same hypoimmunogenic line also showed similar survival in the mice up to 50 days, confirming engineering process did not compromise the iPSC function and its hypo-immunogenicity. Cowen's laboratory also suggested the removal of CIITA in human PSCs but they proposed selective deletion of HLA-A/-B/-C instead of B2M to preserve the expression of HLA-E and HLA-G, the HLA class Ib molecules that retain the tolerance to natural killer (NK) cells (116). They also introduced the expression of PD-L1 (T cell checkpoint inhibitor) and HLA-G in addition to CD47. In their findings, these modifications were able to be protected from the immunosurveillance of T cells, NK cells and macrophages.

This review provides a comprehensive overview of the animal models for cardiac remuscularization study. Successful establishment of the model would need to be confirmed using multiple analyses and imaging such as echocardiography, magnetic resonance imaging, cardiac pressure-volume loop analysis etc. The choice of anaesthesia, analgesia and antibiotic regimen post-surgery is key to increasing the survival of the animal subjects carrying the injured hearts. Nevertheless, the method of choice is based on the experimental needs and objectives. Transplantation of allogeneic cells would require effective immunosuppression to avoid host-vs.-graft rejection of the cells. While the best regimen has not been concluded, the selection of the immunosuppressive strategy is generally aimed toward achieving low toxicity-related side effects, highly efficient immunosuppression, and a high rate of engraftment and survival of the transplanted cells.

Nonetheless, ongoing concerns regarding the incidence of arrhythmias post-CM transplantation (herein refers to engraftment arrhythmias), were possible due to the presence of the nodal cells within the transplanted cardiomyocytes (12, 13, 30, 135). This problem has become the primary impediment to advancing the therapy to clinical trials. Intensive research has been ongoing to decipher pathways that direct chamber-specific cardiomyocyte differentiation to eliminate the presence of nodal cells in the culture and increase the population of ventricular cardiomyocytes in the subsequently transplanted graft. Alternatively, using a pharmacological approach to mitigate engraftment arrhythmias could also be a viable option (22).

In 2014, the European Society of Cardiology Cellular Biology of the Heart Working Group issued a position paper to urge for improving the preclinical assessment of novel cardioprotective therapies. In the statement, the experts attributed the low translatability of laboratory findings into clinics to the lack of rigorous tests during the preclinical animal study (136). One of the shortcomings was the preference over using reductionist cell or rodent models than employing a more integrative large mammal I/R model which simulates clinical reality. We summarize the methodology from the most recent cardiac remuscularization studies using large animals to provide an overview of the differences in reporting between laboratories, and their strategies in establishing MI models, cell source and delivery, as well as post-operative care analgesia and immunosuppression regimen. To increase the reproducibility and transparency of any future in vivo work, adherent to the Animal Research: Reporting In Vivo Experiments (ARRIVE) guideline is urged to facilitate the minimum information and standard required to be included in reporting and publishing animal research experiments (137).

YY, SKT, ZG and JJT conceived the review outline. YY, FFR and JJT wrote the manuscript. YKY, BS, ZG and JJT reviewed and revised the manuscript. All authors contributed to the article and approved the submitted version.