IMPACT: Preclinical studies of cell therapy for human disease

Abstract

The recent scientific statement of the transnational alliance for regenerative therapies in cardiovascular syndromes (TACTICS) provided an overview of the many challenges associated with pre-clinical and clinical studies of stem cell therapy for HF¹ and are providing a series of guidelines and recommendations for moving this field forward^{1, 2}. As part of this series, here we discuss the role of pre-clinical studies designed to advance stem cell therapies for CVD. The quality of this research has improved over the last 10–15 years and overall indicates that cell therapy promotes cardiac repair. However, many issues remain, including inability to provide complete cardiac recovery. Recent studies question the need for intact cells suggesting that harnessing what the cells release is the solution. Here we describe important breakthroughs and current directions in a cell-based approach to alleviating CVD.

I. Introduction to stem cell therapy – landmark preclinical studies/appropriate animal models

Cardiovascular disease (CVD) is the leading cause of mortality worldwide. However, despite improvements in pharmacologic and interventional treatments, 1 in 3 men and 1 in 4 women die within a year of their first myocardial infarction (MI)³. The prevalence of heart failure (HF) and MI require new therapeutic approaches, which must be first tested in animal models to establish safety and therapeutic efficacy, prior to use in humans. Unlike pharmacologic treatments, which primarily manage the disease, stem cell administration promotes the restoration of lost functionality. However, negative outcome trials and the recent debate on the efficacy of the human clinical cell-based therapy in patients with acute MI (AMI)⁴ means that we must continue to find better approaches that will ensure success in human trials.

While myocyte necrosis leads to remodeling post-MI, this effect is secondary to a cascade of cellular changes that appear to be the primary cause of ventricular dilation, hypertrophy and scar formation⁵. In contrast to the age-old paradigm that cardiac myocytes are terminally differentiated, the current consensus is that ~0.5–2% of cardiomyocytes undergo mitosis annually⁶. In infarcted human hearts, myocyte growth becomes enhanced at the border zone after an ischemic event with up to 3–4-fold more dividing myocytes one week post- infarction than in end-stage heart failure⁷. Understanding and enhancing cardiomyocyte proliferation post-MI is a major focus of regenerative medicine.

In early murine studies mobilization of myeloid clonogenic cells from spleen and bone marrow (BM) was observed during wound healing⁸. Later discoveries noted the effects of neovasculogenesis after endothelial progenitor cells (EPCs) mobilized secondary to hind limb ischemia. Rabbits mobilize EPCs specifically from the BM after hind limb ischemia; which was enhanced following GM-CSF administration⁹. These findings paved the way for use of progenitor cell to treat disease. During these early studies, there was no notion of intrinsic self-renewing cardiac cells. In 2003 this paradigm changed; cardiac stem cells that are self-renewing, clonogenic, and multipotent were observed in adult rat hearts ¹⁰. Thus, began the concept that, with some help, the heart could heal itself. The controversy concerned the nature of that help. For many, the answer was which type of stem cell should be used to treat heart disease. The safety, efficacy and fate of each cell line needed further research in animal models to determine not only which model was best to simulate human cardiac response but which of these various cell types should be studied further.

a. Small animal studies

For preclinical development, an appropriate animal model that accurately reflects human pathological conditions is essential. Cell and molecular studies provide important mechanistic data and toxicity studies evaluate candidate drugs¹¹, but a working heart is needed to evaluate and optimize treatments.

New therapies for CVD are usually first evaluated in small animal models (rodents), a model that provides relatively rapid and economical testing and adequate group sizes to ensure sufficient statistical power. Recent technological advances in PET-MRI imaging and echocardiography have improved the assessment of cardiovascular outcomes in rodents¹². Mouse models do have inherent advantages but also some limitations. They can respond very differently than humans to treatment¹³, their hearts beat at 400–600 beats/min and they have a variety of anatomic differences with human hearts (reviewed by Santos et al.¹²). Transgenic and knock-out mice are widely available, making them particularly useful for assessing genetic factors and inducers of cardiovascular diseases. However, genetic changes can alter cardiac morphology, which can limit the advantages of these models¹². Discrepancies between human and mouse embryonic stem cells, including the expression of genes regulating apoptosis, cytokine expression and cell cycle regulation can further limit the relevance of mouse models¹⁴.

Rat heart mass is roughly ten-fold greater than mice, and surgical expertise is less demanding. The rat coronary ligation model was first described in 1979¹⁵, and ligation of the left anterior descending (LAD) coronary artery is the most widely used model for MI. A rat model of MI was instrumental in the evaluation and development of angiotensin-converting enzyme inhibitors^{16, 17} as prelude to clinical trials that resulted in the approval of captopril as a therapeutic intervention for heart failure after MI¹⁸. However, positive rat pre-clinical studies do not necessarily translate to successful clinical trials. Endothelin receptor antagonists such as bosentan, improved survival and hemodynamic characteristics in the rat after MI¹⁹, but failed to show a benefit in humans suffering from heart failure²⁰.

b. Large animal studies

Cardiac repair studies show larger effects in rodents, increased left ventricular ejection fraction (LVEF) up to 20% and normalization of LV function, in contrast to large animal studies (mean LVEF improvement ~5–7%)²¹. This moderate benefit corresponds better to the results of clinical trials, giving realistic insight into the expected benefit of human cell-based cardiac therapies. The presence or absence of collateral coronary circulation is an important factor for choosing an adequate animal model for a particular study. Large animals such as pigs, dogs, or sheep satisfy many of these criteria. Dogs have an extensive collateral coronary circulation, while pigs and sheep have no functionally relevant vascular adaptation system, similar to humans²². Thus, a dog model is suitable for studying vascular adaptation to myocardial ischemia, while pigs and sheep are generally regarded as appropriate to assess the direct myocardial effects of hypoxic injuries.

Initial studies in canine hearts in the late 1970s paved the way to understand myocardial ischemia and the development of HF²³. In a dog MI/reperfusion model, the angiotensin receptor blocker valsartan produced decreased infarct size and increased ejection fraction and improved diastolic function²⁴. Ischemic cardiomyopathies can be simulated in canines through microembolization²⁵, resulting in reduction of LVEF to less than 35%, a model that has been used for evaluation of several drugs for treating heart failure²⁶. The drawback of these studies is that canines have a significantly more complex coronary circulation than humans, such that their maximal oxygen consumption is greater and their degree of reproducible infarct size more variable²⁷. Therefore, despite the advantages of the canine, a better model was still needed. One such model is the sheep, where coronary anatomy is consistent with humans; and there is a lack of significant collateral circulation, allowing for reproducible infarcts. However, sheep carry zoonotic disease, a problem not associated with other large animals and ovine thoracic and gastrointestinal anatomy complicate detailed imaging, so is not ideal for typical transthoracic imaging studies²⁸.

Swine cardiac anatomy compares favorably to that of humans; the large LAD and in most cases, the right coronary artery is dominant, supplying the posterior interventricular septum and the atrio-ventricular node in the vast majority of cases; and there is little collateral flow, all similarities to humans²⁹. Furthermore, adult Göttingen and Yucatan minipigs possess cardiac structure and function comparable to humans, including high mortality associated with large infarcts³⁰. For chronic studies, they remain within a manageable weight range as compared to Yorkshire pigs³⁰. By ligation of the LAD, MI can be induced by either open- or closed-chest methods. Open-chest surgery provides easy access to coronary arteries, visual control of contractility, and facilitates the generation of defined infarction sites and sizes. However, closed-chest procedures avoid the thoracotomy-associated trauma of open-chest surgery³¹ as in humans. Large animal studies are very expensive (10–100x higher than similar small animal studies), limiting the number of animals in a study. A major advantage of large animal studies is the ability to use imaging modalities identical to those for humans, resulting in similar measures and outcome parameters, increasing human relevance but also costs.

This overview of pre-clinical models for regenerative medicine demonstrates that each has inherent advantages and disadvantages. Small animal studies provide an initial indication of the potential of the intervention, and must be further evaluated in large animals whose CV physiology more closely resembles humans. Only when a treatment provides sufficient therapeutic efficacy in large animals should it be moved to clinical trials.

II. Mesenchymal stem cells (MSCs)

Mesenchymal stem cells (MSCs; also called mesenchymal stromal cells), have been the most commonly used stem cell for treating cardiac dysfunction. MSCs are relatively easy to isolate and can be expanded significantly ex vivo. Their immunosuppressive properties and their lack of immunogenicity make them excellent candidates for cell therapy. However, the various methods of cell expansion, isolation and characterization, necessitated a consistent definition of MSCs (Figure 1). The Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy proposed distinct criteria: plastic adherence under standard culture conditions, positive for CD105, CD73 and CD90, and negative for CD45, CD34, CD14 or CD11b, CD79α or CD19 and HLA-DR and the ability to differentiate into osteoblasts, adipocytes and chondroblasts in vitro³². MSCs have an additional therapeutic advantage; they lack MHC class II antigens and not only fail to elicit an immune response (at least initially), but also down regulate host natural killer cells and T-lympocytes³³. Similarly, xenotransplantation of human bone marrow-derived MSCs (BM-MSCs) into murine³⁴ and later porcine³⁵ hearts revealed the ability of (a low percentage of) MSCs to integrate into the myocardium and differentiate into cardiomyocytes.

Figure 1. Procurement and isolation of MSCs.

MSCs isolated from bone marrow and the mononuclear cells isolated by Ficoll density centrifugation. MSCs can be separated from other mononuclear cells by their plastic adherence in culture. From Williams and Hare⁴¹

Cells with similar characteristics to BM-MSCs are found in virtually every tissue³⁶; however, most cardiac studies have used MSCs from BM or a few other tissues (although this trend may be changing) including adipose (Ad)- and umbilical cord blood (UCB)/Wharton’s jelly-derived MSCs. BM and Ad-derived MSCs most consistently differentiate into bone, fat and cartilage, and have the highest capacity for self-renewal³⁷. While circulating MSCs, which are thought to arise from BM, have long been debated as a source for future clinical use, they have produced contradictory results, reducing confidence in them as a source³⁸.

a. Bone marrow-derived MSCs (BM-MSCs)

BM is of particular interest for stem cell therapy because it is a source of a variety of multipotent precursors (MSCs, mononuclear cells, and endothelial progenitor cells; see below). However, stem cells comprise only 1–3% of BM, the majority being lymphocytes; but there are other undesirable cells for cardiac repair: monocytes, pre-adipocytes and osteoblasts³⁹. An important property of MSCs is their ability to home to sites of injury, a property first demonstrated in a baboon lethal radiation model⁴⁰. Murine studies demonstrated that MSCs migrate to ischemic cardiac tissue following intravenous infusion⁴¹. Factors thought to play a role in these migratory and stimulatory properties, include the Stem cell factor/CKit (CD 117) ligand-receptor complex^{42, 43}. Once the ability of MSCs to home was recognized, studies focused on improving cardiac function, which were encouraged by the ability of murine BM-MSCs to become functional cardiomyocytes when treated with 5-azacytidine (5-aza), a demethylase and cytosine analog. MSCs begin beating after 2-weeks of 5-aza exposure, and after three weeks, the beating becomes synchronized⁴⁴.

Overall, most, but not all, studies in small and large animals, for both acute and chronic MI, show improved cardiac structure and function following BM-MSCs administration (see Narita and Suzuki for review⁴⁵). Determining the precise mechanism(s) is difficult, since few cells remain in the myocardium over time. The consensus is that beneficial MSC effects are primarily paracrine, involving not only secretion of growth factors (see Table 1) but also microvesicles and exosomes^{46, 47}. The efficacy of these cells also appears to be influenced by the route of injection⁴⁸.

Table 1.

Paracrine factors secreted by MSCs

Secreted Factor	Function
Pro-angiogenesis
Fibroblast growth factor-2 (FGF-2)	Induces endothelial and smooth muscle proliferation
Fibroblast growth factor-7 (FGF-7)	Induces endothelial cell proliferation
Monocyte chemoattractant protein-1 (MCP-1)	Induces angiogenesis; recruits monocytes
Platelet derived growth factor (PDGF)	Smooth muscle proliferation
Placental growth factor (PlGF)	Promotes angiogenesis
Transforming growth factor-β (TGF-β)	Vessel maturation
Vascular endothelial growth factor (VEGF)	Endothelial cell proliferation, migration, tube formation
Remodeling of extracellular matrix
Metalloproteinase-1 (MMP1)	Loosens matrix; tubule formation
Metalloproteinase-2 (MMP2)	Loosens matrix; tubule formation
Metalloproteinase-9 (MMP9)	Loosens matrix
Plasminogen activator (PA)	Degrades matrix
Tumor necrosis factor -α (TNF-α)	Degrades matrix; cell proliferation
Stem cell proliferation, recruitment, & survival
Basic fibroblast growth factor (bFGF)	Enhance proliferation of endothelial and smooth muscle cells
Granulocyte colony stimulating factor (G-CSF)	Increases proliferation and differentiation of neutrophils
Insulin –like growth factor-1 (IGF-1)	Regulates cell growth and proliferation; inhibits apoptosis
Macrophage colony stimulating factor (M-CSF)	Increases proliferation and differentiation of monocytes
Thymosin-β4 (Tβ4)	Promotes cell migration
Stem cell-derived factor (SDF)	Progenitor cell homing
Secreted frizzled-related protein-1 (SFRP1)	Enhances cell m
Secreted frizzled-related protein-2 (SFRP2)	Inhibit apoptosis; enhance cell development
Immunomodulatory
Heme oxygenase-1(HO1)	Inhibits T cell proliferation
Hepatocyte Growth Factor (HGF)	Inhibits CD4+ T cell proliferation
Indoleamine 2,3-dioxygenase (IDO)	Inhibits innate and adaptive immune cell proliferation
Inducible nitric-oxide synthase (iNOS)	Inhibits inflammation
Interleukin-6 (IL-6)94	Regulates inflammation; VEGF induction
Prostaglandin E2 (PGE2)	Inhibits inflammation

From Williams and Hare⁴¹

1. Small animal studies

Orlic et al. were among the first to show an improvement in cardiac function following intramyocardial injection of (enhanced green fluorescent protein (eGFP)-labeled) lineage (Lin)⁻/c-kit⁺ BM-derived stem cells into female mice in an AMI model. They observed eGFP+ cardiomyocytes, endothelial cells, smooth muscle cells and vascular structures in the infarcted region of the heart, leading to the conclusion that BM-derived cells can engraft and repair myocardium in vivo⁴⁹. Since then small animal studies have reproducibly shown favorable outcomes following AMI. Reduction in infarct size and fibrosis, with enhancement of LVEF and vasculogenesis are among the common findings in the animal trials⁴⁵. Additional benefits include decreased apoptosis, decreased fibrosis, increased vascular endothelial growth factor (VEGF) expression and/or increased regional blood flow in the infarct zone^{48, 50}.

An emerging use for stem cells, particularly BM-MSCs, is in the treatment of idiopathic dilated cardiomyopathy (DCM). Given the lack of distinct areas of hypoperfusion and systolic dysfunction in the absence of coronary disease, it is a challenge to reverse this effect⁵¹. One group established a DCM model in rabbits. BM-MSCs were pre-incubated with 5-aza to induce differentiation into cardiomyocyte-like cells and injected directly into the myocardium resulting in an increase in left ventricular end-systolic pressure (LVESP) and a decrease in LV end-diastolic pressure (LVEDP). There was also an attenuation of myocardial fibrosis with an increase in VEGF⁵². Similarly, in a rat DCM model there was a noticeable decrease in LVEDP and an increase in LVESP⁵³.

Intramyocardial injection of BM-MSCs after AMI produced a significant reduction in fibrosis and a noticeable engraftment of MSCs which differentiated into cardiomyocytes. Infarct size and the number of vascular cells in the myocardial structure were markedly improved⁵⁴. While this study showed high engraftment of MSCs, most studies do not; with engraftment rates ranging from 6–12% of injected cells⁵⁵. In an effort to improve engraftment, Simpson et al. developed an epicardial patch composed of human BM-MSCs and secured it to the acutely infarcted region. One week later, 23% of the cells engrafted into the myocardium. At four weeks, there was less LV dilation and better preservation of wall thickness, but without improvement in ventricular function when compared to controls. The authors attribute the latter finding to the small number of MSCs (1×10⁶) that were initially embedded in the patch⁵⁶.

2. Large animal studies

Both autologous and allogeneic BM-MSC produce beneficial effects in swine acute^{31, 35, 57, 58} and chronic^{30, 59-62} MI models. Early studies highlighted the engraftment and tri-lineage differentiation (CM, endothelium, vascular smooth muscle) of MSCs^{58, 60}. In a porcine model, allogeneic BM- Di-I- and DAPI-labeled MSCs injected into the myocardium after AMI were found in the infarct region 8 weeks later. These cells expressed VEGF and specific cardiomyocyte proteins that suggested an upregulation of vasculogenesis and myocyte differentiation. Imaging and gross inspection showed an increase in LVEF and in subendocardial thickness, and a decrease in scar size in the cell treated group^{57, 63}. While some engraftment occurs, at least in the short term, the consensus is that MSC therapy provides therapeutic efficacy through secretion of growth factors, microvessicles and exosomes (see below). Of particular importance is the observation that intramyocardial injection of BM-MSCs post-MI in Yorkshire pigs increased proliferation of endogenous cardiac stem cells (CSCs)⁵⁸.

In chronic ischemic cardiomyopathy settings, BM-MSC therapy has focused on the ability of MSCs to reduce fibrosis and reverse-remodeling. Reduced myocardial thinning in the infarcted zone was reported in the treatment group following gross inspection. Histologic analysis revealed MSC engraftment up to 6-months post implantation⁶⁴. Schuleri et al., assessed if cell dose is an important parameter. They injected either 20×10⁶ vs. 200×10⁶ autologous cells via an intramyocardial route 12-weeks post-MI. Delayed enhancement showed a decrease in infarct size in the low dose group and a decrease in infarct volume in the high dose group. Myocardial wall thickening was noted in both groups. Contractility was increased in the non-infarcted region in both MSC groups⁶¹; however, the contractility in the infarcted region was increased in only the high dose group. Quevedo et al. similarly tested the efficacy and safety of allogeneic MSCs in a swine model of chronic MI. Animals were injected via a transendocardial route. Twelve weeks post injection, the MSCs had engrafted, vasculogenesis flourished, and myocardial blood flow, LVEF and regional contractility had improved when compared to placebo⁶⁰.

b. Stro-1⁺ BM-MSCs

A subset of BM-MSCs, BM progenitor cells (MPCs) has gained some interest over the past several years. These Stro-1⁺/CD34^– cells are an immature subfraction of BM-MSCs⁶⁵. In addition to their ability to self-renew, they can differentiate into chondrocytes, adipocytes, chondroblasts, and smooth muscle cells. Proponents of this subtype argue that MSCs in general are hard to define completely and this cell subtype can be isolated by immunoselection and has an enhanced ability to replicate and differentiate compared to traditional MSCs⁶⁶. Animal studies with MPCs have produced promising results.

1. Large animal studies

Acutely infarcted sheep received intracoronary injection of allogeneic MPCs. There was a 40% decrease in scar size and a 50% increase in vascular density⁶⁷. Another group used echocardiography to guide the injection of intramyocardial allogeneic MPCs into sheep 4-weeks post-MI. An increase in LVEF, wall thickness and vascular density was reported⁶⁸. In a non-ischemic model, transendocardial administration of ovine allogenic cells produced decreased LVESV, stabilization of LVEF and decreased fibrosis⁶⁹.

c. Adipose-derived stem cells

Adipose-derived stem cells (AdSC) are generated from enzymatic degradation of adipose tissue, which yields the stromal vascular fraction (SVF). The SVF then undergoes adherent culture purification to CD105⁺/CD34⁻ AdSCs. Isolation of adipose tissue (liposuction) is relatively inexpensive and is much less invasive compared to BM aspiration, and yields a large number of cells. AdSCs secrete a similar variety of growth factors as BM-MSC⁷⁰ and ex vivo can be differentiated into cardiomyocytes, endothelial cells, vascular smooth muscle, and evenpacemaker cells⁷⁰. AdSCs, like BM-MSCs are immunoprivileged due to lack of MHC class II and they can improve the function and minimize the immune rejection of endothelial progenitor cells⁷¹. Given those similarities, their ease of isolation and the success of BM-MSCs, there is considerable interest in the potential of AdSCs to improve the function of a failing heart.

1. Small animal studies

While AdSCs and BM-MSCs overall behave similarly, an interesting study was carried out by Rasmussen et al. who isolated AdSC and BM-MSCs from an elderly patient with ischemia. These cells were injected intramyocardially into rats 1 week post-MI. Neither cell type promoted angiogenesis or reduced infarct size, but the AdSCs improved LVEF⁷². This result suggests differences in efficacy as the cells age.

Although not currently being used in human trials, brown adipose tissue may show promise in animal studies. Studies in rats demonstrated that CD29⁺ and CD133⁺ subpopulations of brown adipose tissue reduce infarct size and improve LV function through the induction of cardiomyocyte proliferation^{73, 74}. However more preclinical trials are needed since most of the preclinical studies were conducted on white adipose tissue in mice or rats in an acute setting.

The therapeutic effects of adipose derived-cardiomyogenic cells (Ad-CMG), cells isolated from AdSCs that are CD90⁻, and retain the ability to express cardiac and skeletal muscle proteins, were studied. Three days following ligation of the LAD, mice were injected with Ad-CMG into the coronary artery. Four weeks later, the mice exhibited a reduction in remodeling, increased vasculogenesis, and stabilization of EF⁷⁵.

2. Large animal studies

A porcine model introduced AdSC via coronary artery infusion following AMI. Twenty-eight days later, analysis showed a noticeable decrease in the perfusion defect, an increase in EF, vascular density and wall thickness⁷⁶. Most studies with AdSCs involve AMI models, however one group utilized rabbits to study chronic ischemia. Three weeks post MI the rabbits were injected with AdSCs directly into the infarcted myocardium. AdSCs were associated with a greater vascular density, LVEF and improved EDV 5-weeks post injection compared to controls. However, not all studies demonstrate that AdSCs are very beneficial. Intracoronary administration of allogeneic AdSCs improved perfusion but not LVEF following AMI in a porcine model⁷⁷ but the route of delivery may have affected the efficacy of the cells (see below). Improved perfusion and reduced infarct size was also seen upon administration of AdSCs into humanized pigs 4-weeks post-MI, but only with the highest concentration of AdSCs (4×10⁶ cells/Kg), lower concentrations had no effect on either parameter⁷⁸.

d. Umbilical cord MSCs

Fibroblast-like cells were isolated from the connective tissue of the umbilical cord (Wharton’s Jelly (WJ)) in the early 1990s⁷⁹, and became a source of MSCs. These cells have gained interest for their non-invasive ease of extraction, high MSC yield and short doubling time⁸⁰. There is limited preclinical data on umbilical cord blood and umbilical cord matrix MSCs (UC-MSC/UCM-MSC); however, clinical trials are currently underway. Transcriptional signature comparisons between BM- and UC-MSCs indicates distinct gene expression profiles. For example, UC-MSCs exhibit higher expression of genes associated with cell adhesion, proliferation, immune system functioning neurotrophic support, suggesting that these cells would be better than BM-MSCs for neurodegenerative diseases⁸¹

1. Small animal studies

One group tested a murine model in an acute ischemic setting. After direct intramyocardial injection of hUCM-MSCs, the systolic function was preserved 2 weeks post MI, but there were no differences in infarct size compared to controls, and both groups experienced ventricular wall thinning and dilation. Histologically, there was a lack of cell engraftment into the myocardium, which may indicate that the mechanism of ventricular preservation was due to a paracrine effect⁸². A transgenic mouse model with dilated cardiomyopathy received intramyocardial injections of hUC-MSCs. One month later, the LVEF was increased, the heart weight to body weight ratio decreased by 10%, and chamber dilation reduced in the cell treated group compared to placebo control mice. Histologically, VEGF, IGF-1 were upregulated and vasculogenesis increased, while apoptosis, fibrosis and vacuolization all decreased⁸³. Rat dilated cardiomyopathy was treated with hUCM-MSCs, resulting in reduced fibrosis and cardiac dysfunction. The authors concluded that these cells worked, at least in part, by inhibiting TNF-α and TGF-ß1/Erk1/2 signaling⁸⁴. Roura et al. created a fibrin patch to promote the efficacy of UC-MSC in acutely infarcted mice. The cells were retained within the patch for 4-weeks post implantation. The patch treated group yielded a smaller infarct scar (16 vs. 49%) and increased subjacent myocardial angiogenesis compared to controls⁸⁵. Latifpour et al. injected hUCM-MSCs into the cyanotic region in a rabbit model of AMI. Thirty days later LVEF was significantly improved with evidence of cell engraftment, decreased scar size and chronic inflammatory markers⁸⁶.

2. Large animal studies

Zhang et al. were the first to use hWJ-MSCs in swine after AMI. Six weeks after intramyocardial injection of cells, they observed improved perfusion in the cell treated group, engraftment of injected cells, some of which appeared to have differentiated into cardiomyocytes and vascular endothelial cells. Furthermore, there was an increased proliferation of immature cardiomyocytes expressing c-kit⁺, reduced apoptosis and increased LVEF in the treated group compared to placebo⁸⁷.

Overall, MSCs obtained from a variety of tissues have demonstrated therapeutic efficacy. They tend to work similarly, reducing scar size and immune response, increasing perfusion and improving cardiac function. While MSCs have been studied for a majority of stem cell studies, many other stem cells have promoted cardiac repair.

III. Cardiac stem/progenitor cells (CSCs/CPCs)

A microenvironment or niche exists within the heart that is thought to play a critical role in maintaining stem cells in an undifferentiated state, but releasing them from this “hold” when necessary. Within the niche, stem cells give rise to cardiac progenitor cells, which migrate to sites of myocardial injury in an effort to repair the damage. Unfortunately, these endogenous stem cells quickly become depleted following large infarctions⁸⁸, resulting in incomplete healing and subsequent heart failure. Although the idea of self-renewing cells in the heart was once considered unlikely, this concept is now well accepted and includes myocytes^89-91 as well as stem cells within the myocardium.

a. Small animal studies

The landmark study where c-kit⁺ CSCs were initially identified was conducted in mice. These cells were Lin⁻/c-kit⁺ (CD117; the receptor for stem cell factor (SCF)), previously found in neonatal myocardium⁹², were isolated from the heart, and clonally expanded in culture¹⁰. These Lin⁻/c-kit⁺ cells were injected into the peri-infarct region following MI in rats. The resultant immunohistochemical staining and histological examination showed c-kit⁺ cells self-renew, and act in a clonogenic and multipotent manner to produce cardiomyocytes, smooth muscle cells, and endothelial cells¹⁰. Oskouei et al. directly compared, c-kit⁺ human (h) CSCs hMSCs for cardiac repair in an AMI model in immunodefficient mice. The CSCs produced greater improvement in hemodynamic parameters and were equally able to reduce scar size compared to 30-fold more hMSCs⁹³. Overexpressing Pim1 kinase in c-kit⁺ hCSCs augmented their retention within the myocardium and their therapeutic efficacy in both a mouse⁹⁴ and swine⁹⁵ model of acute MI.

Other notable cardiac stem cells include: Cells expressing stem cell antigen-1 (not found in humans), side population cells (low in c-kit+), and islet-1 transcription factor cells (only found during the neonatal period), none of which to date have been used in clinical trials⁹⁶. Epicardium-derived stem cells, while similarly not yet tested in clinical trials, may be a candidate cell type, but require more preclinical testing. During murine cardiac development, these Wt1⁺ cells develop into functional cardiomyocytes⁹⁷. They are multipotent; resemble MSCs, participate in cardiac development, and likely have the potential to promote myocardial repair⁹⁸. Determining the importance of c-kit+ CSCs for cardiac development and as a cell therapy for heart disease has been fraught with controversy. While some investigators minimize the role of CSCs⁹⁹ recent studies have clarified their origin and differentiation capabilities¹⁰⁰.

b. Large animal studies

In a chronic ischemic swine model, intracoronary administration of c-kit⁺ CSCs into pigs 3-months post MI demonstrated the therapeutic efficacy of these cells. Beginning one month post injection, the LVEF rose in the cell treated group and there was a regional increase in cardiac function. CSCs engrafted and some differentiated into CMs, and vascular structures¹⁰¹.

IV. Cardiospheres and cardiosphere-derived cells (CDCs)

Cardiospheres are a heterogeneous group of stem cells isolated from myocardial biopsies, that form clusters. The cells, initially isolated from human myocardial biopsies, express stem cell-related antigens and some cells spontaneously undergo cardiac differentiation. Upon expansion ex-vivo, these cells are called cardiosphere-derived cells (CDCs)¹⁰². CDCs have been studied as an autologous and allogeneic therapy for MI. Recently it was determined that the CD105⁺/CD90⁻/c-kit⁻ population of CDCs represent the therapeutically active cell fraction¹⁰³.

a. Small animal studies

Coronary infusion of allogeneic CDCs into rats post-MI reduced scar size, and increased cardiac function, myocyte cycling and angiogenesis^{104, 105} Allogeneic CDCs have also proved effective in revitalizing senescent rats¹⁰⁶.

b. Large animal studies

CDC treatment of acute¹⁰⁷ and chronic¹⁰⁸ MI in the pig produces beneficial results. Recently, Gallet et al. demonstrated that CDC therapeutic effects are likely mediated via CDC-derived exosomes¹⁰⁹.

V. Bone marrow mononuclear cells (BM-MNCs)

BM mononuclear cells (BM-MNCs) are a heterogeneous cell population which includes, MSCs, hematopoietic cells, EPCs and others. BM-MNCs are easier to isolate than the component population and have been tested in numerous studies¹¹⁰, but in clinical trials appear to be less effective than MSCs¹¹¹.

a. Small animal studies

Early studies of MI in rats showed that intramyocardial injection BM-MNCs promoted significant vasculogenesis without an associated increase in VEGF or FGF at 2-weeks post injection. However, there was a subsequent decrease of vascularity in the 4-week group, which may have been secondary to the maturation of the scar¹¹². A cryoablation rat model tested mixing BM-MNCs into a fibrin matrix. Eight weeks later, there was a greater enhancement of neovascularization in the cell+matrix group compared to cells alone¹¹³.

b. Large animal studies

Promising results were seen following direct BM-MNCs injection into the peri-infarct zone following LAD ligation. LVEF increased, the perfusion defect markedly decreased and the LVEDV to body weight ratio decreased in the treated group¹¹⁴. Alestalo et al. showed a direct correlation between improvement in LVEF following AMI and the number of retained cells seen on post mortem histologic examination¹¹⁵. Therefore, direct contact and retention of BMCs in ischemic tissue may be critical. Fuchs et al. were the first to establish the safety of transendocardial cell injection in a swine chronic ischemia model. They reported improved perfusion of the ischemic zone and enhancement of wall thickening 4-weeks post MI¹¹⁶, as well as an increase in vascularization of the myocardium and a decrease in scar size in the treated group. However, cardiac function did not change in either group¹¹⁷. A canine model showed the most favorable effects with regard to ventricular function¹¹⁸ and in sheep no difference in function reported in cell-treated animals compared to controls¹¹⁹. Nonetheless, in the preclinical setting, BM-MNCs have produced enhancements to neovascularization and reduced scar size; however, large animal models have shown variable effects in ventricular function.

VI. Hematopoietic and endothelial progenitor cells (EPCs)

The transplantation of BM-derived cells into mice led to the discovery of a hematopoietic source of regenerative cells. Hematopoietic stem cells (HSCs) and EPCs both are CD34⁺/CD133⁺ cells and can transform into myeloid and lymphoid cells or once mobilized from the BM into the blood after tissue injury can subsequently become endothelial cells, which can promote neovascularization¹²⁰.

a. Small animal studies

Murine AMI studies have demonstrated enhanced neovascularization, EF, decreased scar size and differentiation of HSCs into endothelial cells in myocardial tissue^121–123. Manipulating the EPCs provides improved repair potential. Tal et al. compared the ability of EPCs or epigenetically reprogrammed EPCs to differentiate into CMs and promote cardiac repair post-AMI¹²⁴. The unmanipulated EPCs reduced infarct size, reduced LV volume and increased capillary density, but the manipulated (5-aza, valproic acid) EPCs did all three significantly better. Furthermore, the manipulated EPCs engrafted and differentiated into CMs.

b. Large animal studies

EPC conditioned media is cardioprotective following AMI in a swine model. Neutralizing insulin-like growth factor-1 (IGF-1) activity in the media abrogated these effects¹²⁵

VII. Pluripotent stem cells

Transducing mouse fibroblasts with a set of transcription factors (Oct4, Sox2, Klf4, c-MYC) now known as the ‘Yamanaka factors’, produced novel pluripotent cells, induced pluripotent stem cells (iPSCs). iPSCs have surface markers and functional properties similar to embryonic stem cells (ESCs)¹²⁶ and can be differentiated into hormone-responsive beating cardiac cells that mimic ESC-derived cardiomyocytes (ESC-CMs)¹²⁷. Produced from autologous tissue, immune rejection and ethical concerns are no longer an issue. Using iPSC- and ESC-CMs reduces the tumor threat, but the cells remain immature and have limited ability to restore cardiac function¹²⁸.

a. Small animal studies

Although initial animal studies utilized direct injection of iPSCs into the myocardium of mice, resulting in engraftment, improved cardiac function, increased wall thickness and reduced fibrosis¹²⁹. iPSC-CMs engrafted long-term in a rat model of MI but failed to produce beneficial effects¹³⁰

b. Large animal studies

The first studies in a large animal model examined co-injected hiPSCs and hMSCs in acutely infarcted swine. The hiPSCs enhanced vasculogenesis, but the combination of cells increased capillary density to a greater extent, likely secondary to the decreased rates of apoptosis¹³¹. Kawamura et al. placed a sheet of dermal fibroblast-derived hiPSC-CMs over the infarcted area in an ischemic swine model, which produced improved cardiac performance, angiogenesis and an attenuated LV remodeling 8-weeks post implantation¹³². Neither of these two studies reported tumor formation. Both ESC- and iPSC-CMs have been studied in non-human primates. Macaques received 1×10⁹ hESC-CMs 2-weeks post-MI via intramyocardial delivery¹³³. After 3-months, the cells continued to mature (albeit incompletely). Importantly, the cells engrafted, promoted extensive re-muscularization of the infarcted tissue, exhibited regular calcium transients and no evidence of tumors or of cells outside the heart. All of these monkeys exhibited arrhythmias¹³³. Zhu et al. administered hPSC derived cardiovascular progenitors (hPSC-CVPs) into the myocardium of male cynomolgus monkeys 30 minutes post-MI. Cells were present 3 days later but not after 140 days, despite a modified immunosuppressive regimen. Apoptosis was reduced and cardiac function improved by the cells. However, no remuscularization was seen¹³⁴.

VIII. Combination stem cell therapy:

To improve therapeutic efficacy, a new approach is to combine cells. Small and large animal studies have combined progenitor cells. Cells combines with angiogenic/growth factors increased vasculogenesis, cell survival, reduced apoptosis and enhanced cardiac function³³.

a. Small animal studies

Ott et al. were the first to co-inject skeletal myoblasts and BM-MNCs into the myocardium of rats seven days post-infarction. Eight weeks later the combination group demonstrated improved EF/LVEDD/LVEDV, myotube formation and retention of BM-MNCs¹³⁵. Intramyocardial injection of the same combination of cells into canines two weeks post infarction, yielded similar results compared to either cell group individually¹³⁶. Quijada et al. created a fusion between murine MSCs and cardiac progenitor cells (CPCs), termed cardiac-chimeras (CC). They tested the efficacy of CCs in a mouse AMI model compared to the combination of CSC/MSCs or each cell type alone. Four weeks post injection, CC-treated animals showed enhancement of wall thickness. Cardiac function was improved in the CC group at 6 weeks, and in the MSC/CSC group at 18 weeks. Infarct size, engraftment and persistent engraftment were noted in the CC group as compared to MSC/CSC¹³⁷.

b. Large animal studies

The combination of MSCs and CSCs has been studied in swine. Intramyocardial injection of hMSC and/or hCSCs were administered to immunosuppressed swine 14-days post-MI. This combination produced a 2-fold reduction in scar size, 7-fold enhanced engraftment, improved LV compliance and contractility as compared to individual cell types 4 weeks later (Fig. 2). The individual cell types produced significant improvements compared to placebo-treated animals³⁵. In a chronic ischemic, non-immunosuppressed swine model, autologous MSCs±CSCs were administered 3 months post-MI. EF, stroke volume, cardiac output and diastolic strain were all improved in the combination group when compared to MSCs alone. Both cell treated groups significantly improved scar size, wall motion and viable tissue as compared to placebo⁵⁹. A similar study using allogeneic MSCs and/or CSCs again showed that the cell combination produced greater improvements in cardiac structure and function¹³⁸ at least in part by increasing cell proliferation within the myocardium^{59, 138}.

Figure 2. Combination CSC/MSC.

Preload, afterload, and contractility changes after human cardiac stem cells (hCSC) and human mesenchymal stem cell (hMSCs). A. Left ventricular end-diastolic pressure (LVEDP) and (B) end-diastolic volume (EDV) and afterload measured by (C) arterial elastance (Ea). Combination hCSC/hMSC therapy improved contractility as measured by the (D) maximal rate of pressure change during systole (dP/dtmax) and (E) preload recruitable stroke work (PRSW), a preload-independent measure of stroke work. There was no change in (F) systolic elastance (Ees), in any of the groups. All graphs show pre-injection (2 weeks post-MI) vs 4-week post-injection values. Graphs represent mean±SEM. *P<0.05. From Williams et al.³⁵.

IX. Paracrine effects

Considering the reported limited engraftment of transplanted cells, the idea that stem cells secrete factors that activate endogenous cells is particularly attractive¹³⁹. This “secretome” is frequently enhanced by pre-incubation under stressful conditions, such as hypoxia further supports the “paracrine hypothesis”. Table 1 lists some of the potential molecular agents responsible for the paracrine effects. It is likely that a variety of different factors contribute to the regenerative and protective effects. The recruitment of resident stem cells from cardiac tissue or an increased homing of circulating progenitor cells derived from BM are likely enhanced by secreted factors.

a. Cell free medium

Studying stem cell-conditioned medium will help identify the secretome^{140, 141}. For a pig study, MSCs derived from human embryonic stem cells were cultured in serum-free conditions and the cell medium was collected and applied intravenously¹⁴⁰. Three weeks post-MI, pigs treated with conditioned medium showed reduced infract size and preserved cardiac function compared to the control group treated with non-conditioned medium. Capillary density was higher and collagen deposition in border and remote zones were lower in animals receiving the conditioned medium. As mentioned above, (some of) the cardioprotective effects of EPCs appear to be mediated by IGF-1¹²⁵. Percutaneous intramyocardial injection of secretome of apoptotic peripheral blood mononuclear cells (Aposec) decreased infarct size and infarct transmurality measured by cardiac magnetic resonance imaging in a pig model of chronic LV dysfunction¹⁴².

b. Extracellular vesicles/Exosomes

Extracellular vesicles (EVs)/exosomes are membrane bound structures containing a variety of factors including short non-coding nucleic acids microRNAs and proteins. EVs/exosomes have sparked intense interest as the potential mediators of cell-based paracrine effects^143–145. Exosomes purified from MSC-conditioned medium provide cardioprotection in an MI mouse model¹⁴⁶. CSC-derived exosomes recapitulate the major effects of CSCs in both acute and chronic MI mouse models¹⁴⁷. Adamiak et al. recently characterized murine iPSC-derived EVs (iPSC-EVs) which “were enriched in miRNAs and proteins with proangiogenic and cytoprotective properties”¹⁴⁸. Importantly iPSC-EVs provided equal or greater therapeutic efficacy as iPSCs without the potential tumor formation¹⁴⁸. As described above, CDC-derived exosomes produced equivalent levels of cardiac repair as did CDCs in a porcine model¹⁰⁹. These exciting results await comprehensive clinical evaluations.

X. Delivery routes

While the search for the optimal cell type continues, so does the pursuit of the optimal delivery method. The most common routes include: intracoronary, intravenous, and intramyocardial (transendocardial/transepicardial). Despite the various delivery methods, cellular retention remains low³³.

Intracoronary delivery of cells is the most utilized approach in the clinical setting¹⁴⁹. Relatively inexpensive and well tolerated, this approach can be utilized during an acute ischemic event combined with coronary intervention¹⁵⁰. Large animal studies on canines and swine have proven the efficacy of this method^{151, 152}. However arterial obstruction may pose a risk with high cell doses and its application is limited in the chronic ischemia setting, secondary to the diffuse nature of the disease¹⁵³. Intravenous delivery is non-invasive and well tolerated in a swine model¹⁵⁴, yet fairly inefficient, as most of the cells are lost to other organs⁵⁵.

When accompanied by cardiac mapping, the transendocardial approach, although invasive, delivers cells most accurately^{155, 156}. Multiple swine studies have demonstrated improved cardiac function and reduced infarct size with this approach^{57–59, 61, 138}. Caution should be taken in the elderly with thinner myocardium as there have been reports of ventricular perforation and cell clumping in areas of profound ischemia¹⁵⁷.

The transepicardial method requires invasive thoracotomy, but is commonly utilized preclinically³³. Under direct visualization, cells can be precisely injected into viable tissue surrounding the scarred area. Any perforations or hemorrhage can be controlled immediately. This delivery method produced improvements in EF, LVEDV and LVESV compared to placebo in a sheep model 8-weeks post-injection¹⁵⁸. Other large animal studies have reported the benefits of this approach^{35, 64}. Drawbacks include: prolonged post-op recovery, arrhythmias, embolization, and leakage of cells from the injected sites as reported in a swine study¹⁴⁹.

XI. Patches/Biomaterials

As discussed above, generally less than 10% of injected cells are available at the site of injury within a few hours or days after delivery, and few cells actively engraft in the affected tissue. This rapid cell loss represents a major issue not only by limiting engraftment, but also for maintaining paracrine effects, many of which function only locally. Scaffolds/patches have been constructed from various biomaterials (gelatin, Matrigel, and collagen) to mimic the extracellular matrix lost secondary to MI and help retain transplanted cells. These biomaterials include: epicardial patches, self-assembling nanofibers, cell sheets, or injected gels which can be mixed with various cell types and placed on the infarcted region^{159, 160}. Biomaterials need to fulfill a number of (sometimes contradictory) criteria to be effective including: biocompatibility, biodegradability, provide mechanical support, be an appropriate thickness and allow for precise placement¹⁶¹. The advent of 3D printing has expanded the availability and diversity of biomaterials allowing for cell-integration, vascularization and thicker structures¹⁶².

XII. Conclusions and Future Directions

Ranging from in vitro discoveries of the activation and differentiation of stem cells to large animal models mimicking human heart anatomy to culminating in clinical trials, stem cell therapy has been both promising and progressive³³. Here, we discussed promising preclinical studies using a variety of stem cells introduced in different ways to treat a diverse assortment of animals with cardiac diseases. These pre-clinical results show that many types of cells are therapeutic but we must continue to study more cell types and use novel approaches, including combining different types of stem cells, reinjecting cells, improving retention and perhaps using the cell secretome rather than the cell itself.

As with all models, there are caveats associated with current pre-clinical studies. 1) The age of the animal is a crucial factor, with young healthy animals used for most studies. However, in humans, heart failure is primarily associated with aging³. 2) The uniformity of animals of a given strain and lack of co-morbidities, does not match human heterogeneity. 3) Lack of standardized protocols with consistent end points and outcome parameters, as well as multicenter randomized studies with centralized core lab blinded analyses as suggested by the NHLBI sponsored CAESAR consortium¹⁶³ and the Working Group on Cellular Biology of the Heart of the European Society of Cardiology¹⁶⁴. 4) The different injection routes, which appears to influence therapeutic efficacy⁴⁸. Despite these caveats critical interpretation of pre-clinical models is necessary to move regenerative medicine forward.

MSCs have been the most studied cell type. They are widespread, immunomodulatory and immunoevasive and secrete exosomes and growth factors. Despite their demonstrated benefits in acute, chronic and dilated cardiomyopathy, complete recovery using MSCs alone has not occurred. Newer sources of stem cells, including umbilical cord/Wharton’s Jelly have shown potential and a different transcriptome⁸¹. Stem cells isolated from the heart, including c-kit⁺ CSCs and cardiospheres improve cardiac function and scar size in animal models. Pluripotent cells and their derivatives, have similarly shown some promise.

The combination of MSCs and c-kit+ CSCs has proven efficacious in large animal studies^{35, 59, 138} and is being translated to the clinical arena [NCT02503280]. Repeated MSC injections can be more effective than a single administration¹⁶⁵, while Terrovitis et al. followed the transepicardial injection of CSCs with a fibrin based sealant in rats to enhance cellular retention¹⁶⁶. Cell-free systems, particularly MVs/exosomes, with their collection of growth factors, micro RNAs, etc. may represent the next frontier either alone or in combination with cells. We anticipate, that novel pre-clinical approaches will provide more effective treatments and will pave the way for future clinical trials. The current standard of care for heart failure is to prescribe a cocktail of medications for patients to take for the remainder of their lives. The anticipated ability of stem cells to repair a compromised heart means that patients may reduce their medications and live active and healthy lives.

Figure 3. Different administration routes and cell types for treatment of heart disease.

The cell types listed under each delivery method refer only to those referenced (superscripted number) in this review. Figure adapted from Golpanian et al.¹⁶⁷

Non-standard abbreviations

5-aza

5-azacytidine

Ad-CMG

adipose derived-cardiomyogenic cells

AdSC

Adipose-derived stem cells

AMI

acute MI

BM

bone marrow

BM-MSC

bone marrow-derived mesenchymal stem cell

BM-MNC

Bone marrow mononuclear cell

CM

cardiomyocyte

CDC

cardiosphere-derived cell

CVD

Cardiovascular disease

CC

cardiac-chimeras

CPC

cardiac progenitor cell

CSC

Cardiac stem cell

ESC

embryonic stem cell

EPC

endothelial progenitor cell

eGFP

enhanced green fluorescent protein

EV

Extracellular vesicle

HF

heart failure

HSC

Hematopoietic stem cells

DCM

idiopathic dilated cardiomyopathy

iPSC

induced pluripotent stem cell

LAD

left anterior descending

LVEF

left ventricular ejection fraction

LVESP

left ventricular end-systolic pressure

LVEDP

left ventricular end-diastolic pressure

Lin

lineage

MPC

Bone marrow-progenitor cell

MSC

mesenchymal stem cell

MI

myocardial infarction

PSC-CVP

pluripotent stem cell-derived cardiovascular progenitor

SCF

stem cell factor

SVF

stromal vascular fraction

TACTICS

Transnational alliance for regenerative therapies in cardiovascular syndromes

UC

umbilical cord blood

UCM

umbilical cord matrix

VEGF

vascular endothelial growth factor

WJ

Wharton’s Jelly