Present and future of stem cells for cardiovascular therapy

Abstract

In this review we summarize the available evidence regarding the application of stem cell therapy for human cardiovascular repair, going over the principal concepts that will help us understand the present and future of this therapy: first the different types of cells available in clinical practice, second the delivery approaches, and third highlighting the most important clinical studies and their efficacy and safety results.

In addition, we also speculate on the value of current clinical data to gain an insight into the mechanism of stem cell‐based cardiac repair and to design clinical trials in the future.

• Acute myocardial infarction
• cardiovascular
• chronic heart disease
• efficacy
• safety
• stem cells

Introduction

Prognosis for patients with myocardial infarction has improved spectacularly in the last 25 years thanks to the success of strategies that limit necrosis, increase electrical stability of the myocardium and improve ventricular function 1.Since the main underlying cause of postinfarction left ventricle remodeling and dysfunction is irreversible damage to the infarcted wall, the development of treatments aimed at regenerating its muscular and vascular components is now considered a main therapeutic challenge.

The rationale for such an approach includes recent evidence demonstrating the remarkable ability of adult stem cells to produce differentiated cells from embryologically unrelated tissues 2, as well as convincing data supporting the fact that the heart has a potent intrinsic regenerative capacity 3, with the existence of cardiac stem cells able to regenerate myocytes and vasculature after ageing or damage 4–6. In spite of concerns, initial studies suggest that this is also possible in humans 7.

In this review we summarize the available evidence regarding the application of this therapy for human cardiovascular repair, going over the principal concepts that will help us understand the present and future of this therapy; first the different types of cells available in clinical practice, second the delivery approaches, and third highlighting the most important clinical studies and their efficacy and safety results. In addition, we also speculate on the value of current clinical data to gain an insight into the mechanism of stem cell‐based cardiac repair and to design clinical trials in the future.

Concept of stem cell, types and mechanism of action

Stem cells are defined as cells that remain in an early stage of development (immature) capable of dividing indefinitely (self‐renewable) and of giving rise to replica cells (clonogenic) as well as to differentiate to various cell lineages (multipotency).

The most characteristic stem cells are the embryonic stem cells, present in the fertilized oocyte or in the inner cell mass of the blastocyst. As well as in the embryo, stem cells remain in adult tissues (adult stem cells) as skin, skeletal muscle, intestinal mucosa, etc. Adult stem cells stay within the tissues in niches with a specific microenvironment and begin to divide and differentiate into specialized cells of the same germ layer when particular conditions are present. Therefore, two main stem cell sources are available: embryonic and adult stem cells 7. To date, no cardiovascular clinical experience with embryonic cells is available, while the most used adult stem cell sources in clinical trials have been skeletal myoblasts and bone marrow‐derived stem cells (Figure 1).

Figure 1 The most used adult stem cell sources in clinical trials have been skeletal myoblasts(top) and mononuclear bone marrow (bottom) stem cells.

Embryonic stem cells

Embryonic cells are highly proliferative and totipotent, giving rise to all tissues of the adult body. As they are obtained from the inner cell mass of embryos in blastocyst stage, the potential use of human embryonic cells has several problems: the ethical issue, the need of immunosuppression when using allogenic cells for implanting, and its tumorigenic capacity.

Adult stem cells

Skeletal myoblast

These cells are resident satellite stem cells in the muscle and constitute a source of cells with a contractile phenotype. To obtain a sufficient number of cells, myoblasts need to be cultured and expanded for approximately 3 weeks before transplantation 8. Thus, myoblasts have been used only out of the acute clinical setting. It is admitted that myoblasts, which lack electromechanical integration with the host cardiomyocytes, could provide a putative arrhythmogenic substrate 8–10.

Bone marrow‐derived stem cells

Bone marrow mononuclear stem cells

Bone marrow mononuclear stem cells include several cell subpopulations with different morphologic and phenotype characteristics (Table I) 11, 12. Experimental work has shown their capacity to transdifferentiate into myocytes and to promote neovascularization 13. These cells can be used either in the acute or chronic phase of the disease, as their collection is easy and their preparation is not very time‐consuming 7, 14. Only a Ficoll density separation in autologous serum is recommended 15. The great inconvenience is that using whole mononuclear cells it is difficult to analyze the exact role of the different subpopulations in the clinical process.

Table I. Flow cytometry surface expression profile of the different bone marrow cell types used in clinical practice.

Antigen	CD 34	CD 117	CD 133	Mononuclear	Mesenchymal
CD 3	−	−		+	−
CD 14	−	−		+	−
CD 19	−	−		+	−
CD 29	−	−		+	++
CD 34	++	+	+	+	−
CD 44				+	++
CD 45	+	+	+	+	−
CD 56				+	−
CD 90				+	++
CD 105					+
CD 106					+
CD 117	+	++	+	+	−
CD 133	+	+	++	+	−
HLA‐II	+	+	+	+	−

++ Markers positive in the majority of cells; + Markers positive in some cells; − Markers expressed in less than 2% of cells.

Mobilized peripheral blood progenitor cells

Under steady state conditions hematopoietic progenitors are in the bone marrow but can occasionally be detected in circulation. The use of hematopoietic growth factors (for example granulocyte colony‐stimulating factor (G‐CSF)) facilitates the egress of hematopoietic progenitors from bone marrow to peripheral blood. Mobilizing with growth factors could even avoid any direct delivery approach in the coronary artery system or the myocardium.

Selected bone marrow or peripheral blood subpopulations

The expression of some surface antigens CD34, CD133, CD117, characterizes a heterogeneous population of cells including hematopoietic progenitor cells and endothelial progenitor cells. These antigens have been used to isolate specific subpopulations of hematopoietic stem cells. It has been claimed that a subset of CD133+ cells represents endothelial progenitor cells, and the selection method has also been used to enrich the progenitors after mobilization or from bone marrow mononuclear cells.

Bone marrow mesenchymal stem cells

These are a rare population (0.01%–0.001%) of cells from bone marrow stroma, isolated for their capacity of plastic adhesion and their immunophenotype. Mesenchymal stem cells are negative for the markers of hematopoietic stem cells (CD34 and CD133); in fact, these cells lack specific surface markers and have an immunophenotype positive for many adhesion proteins. Bone marrow mesenchymal cells need to be expanded in culture and could only be considered for use in patients with acute myocardial infarction in the allogeneic setting if it is shown that they really are immune‐privileged and will not be rejected.

Adipose tissue‐derived mesenchymal stem cells

Recently, the existence of a population of cells has been seen, located in the adipose tissue, which are able to differentiate into multiple cell lineages. The phenotype of adipose stem cells is also similar to that of bone marrow mesenchymal cells. What makes this source so attractive is the possibility to extract large volumes of tissue and cells avoiding culture.

Cardiac resident stem cells

Several reports have demonstrated the existence of cycling ventricular myocytes in both the normal and the pathologic adult heart, challenging the classic dogma which stated that the adult mammalian heart is an organ without regenerative capacity. Their capability to differentiate into cardiomyocytes or vascular lineages suggests that these cells could be used for cardiac tissue repair 16. To date, several subpopulations have been identified: c‐kit cells, stem cell antigen‐1 (sca‐1), side population (SP) cells, cardiospheres, and Isl‐1⁺ cells 16. In spite of their presence in adult heart, spontaneous total recovery after myocardial injury does not occur since a very insufficient number of cardiac stem cells is available. However, they could potentially be expanded in culture or stimulated with growth factors.

Although different types of stem cells can be used in clinical practice, we do not know at present which cell subtype is the more efficient. Each subtype has advantages and pitfalls, but one important issue is the time needed for their preparation: if it is time‐consuming they cannot be used in the acute clinical setting. Moreover, in conditions where early microvascular reperfusion or chronic ischemia prevails, the angiogenic potential of bone marrow‐derived cells may be of high priority. However, in patients where additional restoration of contractile function is the clinical goal, such as those with ischemic cardiomyopathy and heart failure or early post infarction when blood flow has been restored but cardiocytes have died, delivering cells with contractile potential seems a more reasonable approach; under these conditions, skeletal myoblasts, or any progenitor cells driven down a muscle lineage appear a better first choice.

Key messages

• The principal differences between the available clinical studies have been the type of patients studied, the type of cell used, the delivery approach, and the method of measurement.

• Two main stem cell sources are available: embryonic and adult stem cells. To date, no cardiovascular clinical experience with embryonic cells is available, while the currently used adult stem cell sources in clinical trials have been skeletal myoblasts and bone marrow‐derived stem cells.

• Clinical data show that this therapy is a feasible and safe procedure that seems to diminish significantly the extent of left ventricular remodeling, promote recovery of both global and segmental cardiac function, and improve myocardial perfusion. At this point experts advocate conducting intermediate‐size, randomized controlled trials to establish the effects of stem cell therapy using surrogate end points, not forgetting safety concerns.

Mechanism of action

According to the classical view, adult stem or progenitor cells seemed to have more limited potential than embryonic cells, being able to generate cells only within the same lineage. However, in recent years several studies have shown that circumstances such as tissue injury may act as a trigger for adult stem cell differentiation into specialized cells from embryological unrelated tissues. When metaplasia (conversion of one cell type into another) happens to a differentiated cell the process is named transdifferentiation. Bone marrow‐derived stem cells (hematopoietic and mesenchymal) appear as a representative example of plasticity given their outstanding capability to give rise to a wide range of cell types from different organs. Indeed, some authors suggest that the mechanism by which partially or totally differentiated cells change their phenotype might be a process of in vivo stem cell fusion with differentiated host cells 17. To date it is unknown whether transdifferentiation or cell fusion is the main mechanism of cardiac repair. Nevertheless, the positive influence of stem cell therapy could hardly be explained only thanks to cell replacement, as the number of cells (approximately 3%) that remain in the cardiac tissue is limited and insufficient in hearts with large injuries 18. A paracrine effect of extracardiac progenitors on cardiac resident stem cells might have something to do with the benefits 19. In this way, such mechanisms seem to be complementary rather than opposite, and their understanding becomes the main challenge in the future. Therefore, although regeneration does exist, it is yet to be determined whether cardiac regeneration or repair is responsible for the benefit.

Delivery approach strategies

The aim of any route of stem cell delivery is to implant a sufficient number of cells to gain the maximum possible retention in the region of the myocardium to be treated. The local host is a crucial determinant of the cellular retention since it determines cellular survival in the short term, cellular adhesion and transmigration across the vascular wall, and cellular tissue invasion.

To date, two principal routes of administration have been used: the transvascular approach and a direct injection in the ventricular wall.

Transvascular approaches

Intracoronary infusion

Obviously, to administer the cells via the coronary arteries they have to be patent. Thus, an intracoronary percutaneous revascularization of the culprit artery is a prerequisite to ensure a vascular access through which the stem cells will be given. An over‐the‐wire angioplasty balloon is used to deliver the cells into the coronary artery. Alternating periods of occlusion of the artery and infusion of the cells (2–3 minutes) with periods of reperfusion (1–3 minutes) are used until the total dose is given. Currently, intracoronary administration of hematopoietic progenitors represents the best‐tried method of delivery in the acute setting.

Intravenous infusion

In animal models the intravenous delivery of progenitor and mesenchymal cells has shown to improve left ventricular function after an acute coronary syndrome 4, 20. However, the application in clinical studies is limited considering seeding in organs other than the heart. In a recent study myocardial homing was not observed after intravenous administration when compared to intracoronary infusion 21.

Mobilization with growth factors

The pharmacological mobilization of bone marrow stem cells with growth colony‐stimulating factor is a very attractive alternative considering its easy administration (subcutaneously) avoiding any invasive procedure. The process by which this mobilization takes place is not completely understood, but it is well known that it is mediated by adhesion molecules and through trophic chemokines 22.

Direct injection in the ventricular wall

This approach should be preferred when the coronary vascularization avoids intracoronary infusion. However, the direct injection in the ischemic myocardium could create cell islets with scarce blood flow that could limit cell survival 23. This technique allows big‐size cells to be injected (mesenchymal and myoblasts) that could produce microembolisms with the intracoronary delivery 24.

Transendocardial injection

The electromechanical mapping of the endocardial surface using the NOGA® system (Cordis Corporation's Biologics Delivery Systems) allows viable, ischemic, and scar myocardium to be defined before the injection through a catheter positioned in the left ventricle. This approach has been used in patients with chronic heart disease.

Transepicardial injection

This approach is used in patients going to coronary artery bypass grafting surgery (CABG). The injection procedure is generally performed after all the graft sutures have been done, before removing the extracorporeal circulation, and while the heart is initiating spontaneous heartbeat. The injection area is often easily identifiable as a patchy fibrotic area with a dyskinetic or akinetic wall motion assessed by the specific imaging techniques. This approach allows a large amount of stem cells to be delivered, but there are myocardial regions not accessible (i.e. the interventricular septum).

Transvenous injection

This approach has been experimented in a pilot study in patients with ischemic cardiomyopathy using myoblasts and guided with intravascular ultrasound. In contrast to the transepicardial injection, with this method cells are injected in parallel to the interventricular septum.

Stem cell delivery is a critical step toward the optimization of the therapy and techniques to find out the destiny of stem cells will be crucial if we want to understand the mechanisms of action and to improve the short‐ and long‐term efficacy of this treatment. Stem cells labeled with various substances have been tracked in animal models by means of magnetic resonance, positron emission tomography, single‐photon emission computed tomography, and bioluminescent imaging 25. However, stem cells infused into humans have been tracked in only one study to date 21. Unfractionated cells and CD34+ cells were labeled with 18F‐fluorodeoxyglucose and infused via an antecubital vein or a coronary artery after a first myocardial infarction in nine patients. After intracoronary transfer only 1.3%–2.6% of unselected bone marrow stem cells and 14%–39% of CD34 cells were detected in the infarcted myocardium; the remaining activity was found primarily in liver and spleen. Furthermore, only background activity was detected in the infarcted myocardium after intravenous transfer. Integration of cell imaging into studies of cardiac cell therapy is necessary to further growth of the field towards optimizing routes of stem cell delivery.

Clinical studies—efficacy and safety

The evidence that stem cells may reconstitute necrotic myocardium and improve cardiac function in animals has led to an initiation of clinical studies, which have been mainly focused on addressing the feasibility and the safety of this therapy. The vast majority of studies have been performed in patients with acute myocardial infarction (Table II). Other important groups of studies have been performed in patients with chronic ischemic cardiomyopathy (Tables III and IV). Thus, the principal differences between these studies were the clinical scenario, type of cell employed, the delivery approach, and the method of efficacy measurement (Figure 2).

Table II. Cell therapy clinical trials in humans after acute myocardial infarction.

Author	Design ^a and groups	Dose	LVEF (%) ^b		Timing ^c	F‐U ^d	Results
			Baseline	F‐U
Strauer 42, 43	Phase I				5–9	3	Cell‐treated: ↓ infarct size and ↑ contractility of infarcted area.
	i.c. BMMC (10 p.)	2.8±2.2×10^7	57±8	62±10
	NR control (10 p.)	No placebo	60±7	64±7
Assmus 44, 45	Phase I				5±2	4, 12	Both types of cells: ↑ LVEF, ↓ ESV, infarct size reduction and absence of reactive hypertrophy. Sustained benefits at 12‐month F‐U.
	i.c. BMMC (29 p.)	1.6±1.2×10^7	51±10	59±10
	i.c. CPC (30 p.)	21.3±7.5×10^7	49±10	57±10
	NR control (11 p.)	No placebo	51±10	54±8
Chen 46	Phase II				18	6	Cell‐treated: ↑ LVEF, ↓ ESV and EDV, ↑ perfusion and contractility of infarcted area.
	i.c. MSC (34 p.)	4.8–6.0×10^10	49±9	67±3
	R control (35 p.)	Standard saline	48±10	54±5
Kuethe 47	Phase I				6	12	No changes in LVEF and contractility of infarcted area.
	i.c. BMMC (5 p.)	3.9±2.3×10^7	42±8	46±7
Fernández‐Avilés 15, 48	Phase I				14±6	6	Cell‐treated: ↑ LVEF, ↓ ESV, ↑ contractility and thickness of infarcted wall.
	i.c. BMMC (20 p.)	7.8±4.1×10^7	51±7	57±10
	NR control (13 p.)	No placebo	50±7	52±7
Wollert 49, 50	Phase II				6±1	18	Cell‐treated: acceleration of systolic function recovery (6‐month F‐U: ↑ LVEF, ↑ regional contractility), no difference with controls at long term.
	i.c. BMNC (30 p.)	246±94×10^7	50±10	56±15
	R control (30 p.)	No placebo	51±9	54±13
Bartunek 36	Phase I				12±1	4	Cell‐treated: ↑ LVEF, ↓ ESV, ↑ perfusion and contractility of infarcted area. Besides, ↑ incidence of coronary events, ISR, de novo lesions.
	i.c. CD133+(19 p.)	1.3±0.2×10^7	45±3	52±4
	NR control (16 p.)	No placebo	44±3	49±4
Katritsis 51	Phase I				8–1,560	4	Recent AMI cell‐treated patients: ↑ regional contractility, viability, and reversible ischemia.
	i.c. MSC+EPC (11 p., 6 of them with recent AMI)	2–4×10^6 e	40±9	41±10
	NR control (11 p.)	No placebo	42±11	42±10
Ge 52	Phase II, double‐blinded				0–1	6	Cell‐treated: ↑ LVEF, ↑ regional perfusion. Left ventricular dilatation in controls.
	i.c. BMMC (10 p.)	4×10^7	54±9	59±10
	R control (10 p.)	Placebo	58±8	56±4
Janssens 53	Phase II, double‐blinded				2	4	Cell‐treated: ↑ perfusion and contractility of infarcted area. No difference in LVEF recovery.
	i.c. BMMC (33 p.)	172±72×10^6	49±7	52±9
	R control (34 p.)	Placebo	47±8	49±11
Lunde 27	Phase II				5–8	6	No changes in LVEF assessed by 3 methods (MRI, SPECT, 2D‐echo).
	i.c. BMMC (50 p.)	87±47×10^6	55±14	56±15
	R control (50 p.)	No placebo	54±12	58±11
Schächinger 26, 54	Phase II, double‐blinded				3–6	4	↑ LVEF and contractility of infarcted area, ↓ combined end point: death, reinfarction, and rehospitalization.
	i.c. BMMC (101 p.)	236±174×10^6	48±9	54±10
	R control (103 p.)	Placebo	47±10	50±13
Kuethe 55, 56	Phase I				2 (for 7±1 days)	3	G‐CSF group: ↑ perfusion and contractility of infarcted area; Both groups: ↑ LVEF.
	G‐CSF (14 p.)	10 µg/kg/d	40±11	48±13
	NR control (9 p.)	No placebo	40±13	43±13
Valgimigli 57	Phase I, single‐blinded				2±3 (for 4 days)	6	Nonsignificant trend in treated patients to ↑ LVEF and ↓ EDV.
	G‐CSF sc. (10 p.)	5 µg/kg/d	NA	NA ^f
	R control (10 p.)	Placebo	NA	NA
Suárez de Lezo 58	Phase I				5 (for 10 days)	3	↑ Contractility of infarcted area; spontaneous spleen rupture on day 8 on G‐CSF (1 p.).
	G‐CSF sc. (13 p.)	10 µg/kg/d	40±7	46±15
Ince 38, 59	Phase II				<1 (for 6 days)	4, 12	G‐CSF group: ↑ LVEF, perfusion, thickness, and contractility of infarcted wall.
	G‐CSF sc. (25 p.)	10 µg/kg/d	48±4	54±8
	R control (25 p.)	No placebo	47±5	43±5
Jorgensen 29, 60	Phase II, double‐blinded				1–2 (for 6 days)	6	No differences in LVEF, volumes, ventricular mass, and systolic thickening between groups. No neointimal hyperplasia with G‐CSF.
	G‐CSF sc. (39 p.)	10 µg/kg/d	51±15	60 ^g
	R control (39 p.)	Placebo	56±10	64
Zohlnhöfer 30	Phase II, double‐blinded				5 (for 5 days)	4, 6	No differences in LVEF, volumes, and perfusion between groups.
	G‐CSF sc. (56 p.)	10 µg/kg/d	51±8	52±8
	R control (58 p.)	Placebo	49±9	51±9
Engelmann 32	Phase II, double‐blinded				0–7 (for 5 days)	3	G‐CSF group: ↑ perfusion in infarcted area at 1‐month follow‐up. No differences in LVEF and volumes between groups.
	G‐CSF sc. (23 p.)	10 µg/kg/d	41±12	47±12
	R control (21 p.)	Placebo	44±9	50±12
Ellis 31	Phase II double‐blinded				<2 (for 5 days)	1	No differences in LVEF, regional contractility, volumes, and infarct wall thickness between groups.
	G‐CSF 10 µg/kg/d (6 p.)	10 µg/kg/d	37±8	39±7
	G‐CSF 5 µg/kg/d (6 p.)	5 µg/kg/d	42±8	34±5
	R control (6 p.)	Placebo	34±2	41±10
Kang 37, 39	Phase II				4—6 (G‐CSF for 4 days) ^h	6	Combined treatment group: ↑ LVEF, ↓ EDV, ↑ perfusion of infarcted area, ↑ exercise capacity. Study prematurely stopped due to high rate of in‐stent restenosis (G‐CSF groups).
	G‐CSF sc.+i.c. CPC (7 p.)	10 µg/kg/d	49±9	55±7
	G‐CSF sc. (3 p.)	CPC: 100×10^7	45±14	44±16
	R control (2 p.)	No placebo	40±9	NA
Kang 61	Phase II				7±1 (G‐CSF for 3 days) ^i	6	Recent AMI cell‐treated patients: ↑ LVEF, ↓ ESV.
	G‐CSF sc.+i.c. CPC (25 p.)	10 µg/kg/d	52±10	57±9
		CPC: >7×10^6
	R control (25 p.)	No placebo	53±13	53±12
Steinwender 62	Phase I				6—8 (G‐CSF for 4–6 days) ^k	6	↑ LVEF and regional contractility. High incidence of ISR and reinfarction.
	G‐CSF sc.+i.c. CPC (20 p.)	10 µg/kg/d and CPC ^j	46±8	54±11

Values are mean±standard deviation or range. LVEF = left ventricular ejection fraction; F‐U = follow‐up; BMMC = bone marrow mononuclear cells; BMNC = bone marrow‐nucleated cells; G‐CSF = granulocyte colony‐stimulating factor; NR control = nonrandomized control group; R control = randomized control group; CPC = circulating progenitor cells; ESV = end‐systolic volume; EDV = end‐diastolic volume; MSC = mesenchymal stem cells; AMI = acute myocardial infarction; MRI = magnetic resonance imaging; SPECT = single photon emission computerized tomography; 2D‐echo = two‐dimensional echocardiography; ISR = in‐stent restenosis; sc. = subcutaneous administration; i.c. = intracoronary administration; p. = patients; NA = not available; NS = no significant; PCI = percutaneous coronary intervention.

^a All studies were open label unless stated otherwise.

^b LVEF data showed in the table were calculated or measured by left ventricular angiography 26, 36, 42–47, 54, 58, 62, 2‐dimensional echocardiography 31, 38, 52, 59, radionuclide ventriculography 51, 55, 56, cardiac magnetic resonance 15, 27, 30, 49, 50, 53, 60, 61, and gated‐SPECT 37, 57.

^c Timing indicates the number of days from the onset of symptoms to the administration of the treatment (and its duration in the case of G‐CSF).

^d F‐U indicates the number of months from treatment administration to follow‐up imaging assessment.

^e Including mesenchymal stem cells (66±19%) and endothelial progenitor cells (28±11%).

^f Only relative changes of LVEF at 6‐month follow‐up were reported; no significant differences were found between G‐CSF patients and controls.

^g Absolute changes in LVEF reported (8.5% and 8% in G‐CSF and control group respectively; P = NS), standard deviation not mentioned at follow‐up.

^h Cell transplantation was carried out immediately after G‐CSF completion (10 µg/kg/d for 4 days prior to stent implant) and PCI‐stenting.

ⁱ G‐CSF treatment (10 µg/kg/d for 3 days) started after successful PCI (7±1 days postinfarction); cell transplantation was carried out immediately after G‐CSF completion.

^j CPC obtained by apheresis included 49±37×10⁶ CD34+ cells.

^k G‐CSF treatment for 5.5±1.2 days was initiated 2 days after the onset of symptoms; intracoronary transplantation of progenitors collected by apheresis was carried out immediately after C‐CSF completion.

Table III. Clinical trials with direct intramyocardial stem cell injection in humans with chronic ischemic heart disease.

Author	Type of patients (n)	Timing (months)	Cell type and dose ^a	Delivery method	LVEF (%)		Results
					Baseline	F‐U
Menasché and Hagege 8, 63	ICM, OMI, and indication for CABG in remote reas (10 p.)	3–229	Myoblasts 871±62×10^6 (86%)	TEP ^b	24±1	32±1	At 11±5 months of F‐U: ↑ LVEF ^c and systolic thickening. NYHA functional class improvement. 4 cases with delayed SVT (ICD).
Herreros and Gavira 34, 64	ICM, OMI, and indication or CABG in remote areas (11 p.)	3–168	Myoblasts 221±100×10^6 (66%)	TEP	36±2	54±5	At 3 and 12 months of F‐U: ↑ LVEF, regional contractility, and viability.^d
	Historical NR control (14 p.)
Siminiak 65	ICM, OMI, and indication for CABG in remote areas (10 p.)	4–108	Myoblasts 0.4–50×10^6 (65%)	TEP	35 (25–40) ^e	42 (29–47) ^e	At 12 months of F‐U: ↑ LVEF and regional contractility. SVT in 2 patients, controlled with amiodarone.
Cachques 66	ICM, OMI, and indication for CABG in remote areas (20 p.)	NA	Myoblasts 300±20×10^6 (>70%)	TEP	28±3	52±5	At 6 months of F‐U: ↑ LVEF, regional contractility, and viability.
Dib 67	ICM, OMI, and indication for CABG (24 p.), or for LVAD (6 p.) ^f	NA	Myoblasts Escalating dose (1, 3, 10, 30×10^7) (79%) ^g	TEP	28±9	36±11	At 24 months of F‐U: ↑ LVEF, ↑ viability, ↓ LV volumes. NYHA functional class improvement. Delayed episodes of NSVT in 3 patients (ICD implanted).
Stamm 68, 69	Recent MI and indication for CABG (12 p.)	0.6–3	CD133+ escalating dose (0.12–5×10^6)	TEP	40±9	49±6	At 2 weeks of F‐U: ↑ LVEF and perfusion, ↓ LV volumes. NYHA functional class improvement.^h
Mocini 70	Recent MI and indication for CABG in remote areas (18 p.)	1–6	BMMC (292±232×10^6)	TEP	46±6	51±9	At 3 months of F‐U: ↑ LVEF and regional contractility when compared to controls.
	NR control (18 p.)
Patel 71	Phase II trial, single‐blinded.	NA ^i	CD34+ (22×10^6)	TEP	29±4	46±2	At 6 months of F‐U: ↑ LVEF and NYHA functional class improvement.^d
	ICM, symptomatic HF, and indication for CABG in remote areas (10 p.)
	R control (10 p.)
Archundia 72	ICM, OMI and symptomatic HF (5 p.)	13–19	CPC (400×10^6) ^j	TEP	38–45	55–60	At 7 months of F‐U: ↑ LVEF, perfusion, and systolic thickening of infarcted area, ↓ LV volumes in treated patients.
	NR control (10 p.)
Galinanes 73	OMI and indication for CABG in remote areas due to angina (14 p.)	23±17	BMMC (32±4×10^6 per segment) ^k	TEP	47±4	48±4	At 10 months of F‐U: NYHA and Canadian functional class improvements, ↑ contractile reserve and WMSI whenever CABG and cell injection were combined.
Hamano and Li 74, 75	Severe ischemic heart disease with indication for CABG and ungraftable area/s (5 p.)	NA	BMMC (30‐200×10^7)	TEP	NA	NA	At 6 months of F‐U: ↑ perfusion in treated area in 3 of the 5 patients. Safety proven at 1 year of F‐U.
Ozbaran 76	HF and ICM not suitable for CABG (6 p.)	NA	CPC (5–64×10^9) ^j	TEP+i.c.	21±3	28±10	At 4 months: increase in life quality and NYHA functional class.
Pompilio 77	Ischemia not suitable for standard revascularization or recent MI with indication for CABG in remote areas (4 p.)	NA	CD133+ (13–200×10^6) ^l	TEP	NA	NA	Safety of this strategy proven.
Siminiak 78	ICM, OMI, and symptomatic HF (10 p.)	5–96	Myoblasts 17–106×10^6 (>65%)	Transcoronary venous ^m	40±7	43±8	At 6 months of F‐U: ↑ LVEF in 66% of the patients. NYHA functional class improvement. SVT in 1 p. with ICD previously implanted.
Smits 79	ICM, OMI, and symptomatic HF (5 p.)	24–132	Myoblasts 196±105×10^6 (55%)	TEND ^n	36±11	45±8	At 6 months of F‐U: ↑LVEF, contractility, and systolic thickening of infarcted area. NSVT in 1 patient (ICD implanted).
Ince 80	ICM, OMI, and symptomatic HF (6 p.).	79±59	Myoblasts 210±150×10^6 (70%)	TEND ^o	24±7	32±10	^p At 12 months of F‐U: ↑ LVEF, contractility, and contractile reserve of infarcted area. NYHA functional class improvement. SVT episodes in 2 p. with ICD previously implanted.
	NR control (6 p.)
Perin 81, 82	ICM, OMI, and symptomatic HF with no option for conventional revascularization (14 p.)	>3	BMMC (26±6×10^6)	TEND ^n	30±6	36±8	At 4 and 12 months of F‐U: ↑ contractility and perfusion of infarcted area, and NYHA and Canadian functional class improvement. Sudden cardiac death (1 p.) at 14 weeks of F‐U.
	NR control (7 p.)
Silva 83	ICM, OMI, and HF with no option for conventional revascularization awaiting cardiac transplantation (5 p.)	NA	BMMC (30×10^6)	TEND ^n	30±8	28±12	At 6 months of F‐U: ↑ myocardial volume oxygen consumption and exercise capacity, and 4 of the 5 patients were no longer eligible for cardiac transplantation.
Tse 84, 85	Refractory stable angina (8 p.)	NA	BMMC (dose NA)	TEND ^n	58±11	61±15	At 3 months of F‐U: ↑ contractility, perfusion, and thickening of treated area. Symptomatic improvement. Safety proven at very long term.^q
Fuchs 86, 87	Refractory stable angina with no option for conventional revascularization (27 p.)	NA	BMMC (67±65×10^6)	TEND ^n	48±9	50±7	At 3 and 12 months of F‐U: ↑ exercise capacity, ↓ angina score, and ↓ stress‐induced ischemia in treated areas.
Briguori 88	Refractory stable angina with no option for conventional revascularization (10 p.)	NA	BMNC (127–350×10^6)	TEND ^r	53±10	57±16	At 12 months of F‐U: ↑ perfusion, improvement in the physical limitation, angina frequency, angina stability scores.

All studies mentioned are open label phase I trials unless stated otherwise. Timing indicates the number of months from myocardial infarction to the administration of cell therapy; delivery method is shown as TEP (transepicardial injection), TEND (transendocardial injection), transcoronary venous approach, and i.c. (intracoronary). LVEF indicates left ventricular ejection fraction, and data showed in the table were calculated or measured by left ventricular angiography 79, 2‐dimensional echocardiography 8, 34, 63–65, 67–73, 76, 78, 80–85, radionuclide ventriculography 66, and cardiac magnetic resonance 86, 87.

F‐U = follow‐up; ICM = ischemic cardiomyopathy; OMI = old myocardial infarction; HF = heart failure; CABG = coronary artery bypass graft; LVAD = left ventricular assist device; NR control = nonrandomized control group; p. = patients; SVT = sustained ventricular tachycardia; NSVT = nonsustained ventricular tachycardia; ICD = implantable cardioverter defibrillator; LV = left ventricular; BMMC = bone marrow mononuclear cells; BMNC = bone marrow‐nucleated cells; G‐CSF = granulocyte colony‐stimulating factor; CPC = circulating progenitor cells; MI = myocardial infarction; NA = not available; NYHA = New York Heart Association; WMSI = Wall Motion Systolic Index.

^a Values are mean dose±standard deviation or range or escalating dose as indicated; in the case of studies using myoblasts, the mean percentage of myoblasts injected is shown between parentheses.

^b Unique study of transepicardial injection of myoblasts during CABG in which the grafted scar was not bypassed.

^c Sustained benefit at long‐term follow‐up (median time 52 months).

^d Significant benefit when compared to control group of isolated CABG patients.

^e Mean LVEF (and range).

^f As a bridge to heart transplantation.

^g Cell dose in LVAD group was 30×10⁷, except for 1 patient, who received only 2.2×10⁶ cells.

^h Sustained benefit more than 1 year afterwards in 5 patients.

ⁱ 90% of treated patients had previous myocardial infarction.

^j Circulating progenitor cells obtained by apheresis after stimulation with G‐CSF.

^k Including 0.6±0.1×10⁶ CD34+/CD117+ cells/segment.

^l Circulating progenitor cells obtained by apheresis after stimulation with Hematopoietic Growth Factor Lanogastrim.

^m TransAccess (TransVascular Inc.) catheter system.

ⁿ Guided by electromechanical mapping with NOGA catheter system (Biosense‐Webster).

^o MyoCath injection catheter system (Bioheart).

^p Significant improvements at follow‐up versus baseline and versus controls.

^q Mean follow‐up of 44±10 months in 12 patients.

^r Guided by electromechanical mapping with NOGA catheter system and performed with percutaneous injection using a MYOSTAR injection catheter (Biosense Webster).

Table IV. Clinical trials with intracoronary administration and/or intravascular mobilization of stem cells in humans with chronic ischemic heart disease.

Author	Design ^a and groups	Dose	LVEF (%) ^b		Timing ^c	F‐U ^d	Results
			Baseline	F‐U
Kuethe 89	Phase I				16±6	3, 12	No changes in LVEF, regional contractility, microvascular function, and peak oxygen uptake at F‐U.
	i.c. BMMC (5 p.)	29±9×10^6	29±8	34±8
Strauer 90	Phase I				27±31 (5–102)	3	Cell‐treated group: ↑ LVEF, contractility, and perfusion of infarcted area.
	i.c. BMMC (18 p.)	90×10^6	52±9	60±7
	NR control (18 p.)	No placebo	51±10	52±10
Katritsis 51	Phase I				0.2–52	4	OMI cell‐treated patients: ↑ reversible ischemia in treated area.
	i.c. MSC+EPC (11 p., 5 of them with OMI)	2–4×10^6 e	40±9	41±10
	NR control (11 p.)	No placebo	42±11	42±10
Blatt 91	Phase I				58±2 (12–120)	4	NYHA functional class improvement. In hibernating segments also ↑ regional contractility and contractile reserve.
	i.c. BMNC (6 p.) ^f	13–65×10^8	25±7	28±8
Assmus 33	Phase II				>3 (3–300)	3	BMC‐treated patients: ↑ LVEF and contractility of treated area, also in a crossover phase.
	i.c. BMMC (28 p.)	205±110×10^6	41±11	43±10
	i.c. CPC (24 p.)	22±10×10^6	39±10	39±10
	R control (23 p.)	No placebo	43±13	42±13
Erbs 92	Phase II, double‐blinded				8±3^g	3	Cell‐treated group: ↑ LVEF, contractility, and perfusion of treated area, and improvement of microvascular function.
	G‐CSF sc.+i.c. CPC (13 p.)	G‐CSF 300 µg/d for 4 days and CPC: 69±14×10^6	52±4	59±3
	R control (13 p.)	Placebo	56±3	56±3
Kang 61	Phase II				17±17 ^h	6	OMI cell‐treated patients showed microvascular function improvement. No changes in LVEF or LV dimensions.
	G‐CSF sc.+i.c. CPC (16 p.)	10 µg/kg/d for 3 days and CPC (>7×10^6)	49±13	49±13
	R control (16 p.)	No placebo	45±10	45±11
Seiler 93	Phase II, double‐blinded				NA ^i	0.5	GM‐CSF group improved collateral flow.
	GM‐CSF sc. (10 p.)	10 µg/kg/48h	65±10	NA
	R control (11 p.)	Placebo	58±18	NA
Wang 94	Phase I				NA	2	G‐CSF group: ↓ nitroglycerin consumption, angina attacks. No effect on perfusion or LVEF.
	G‐CSF sc. (13 p.)	5 µg/kg/d for 6 days	48±10	44±12
	NR control (16 p.)	No placebo	46±11	43±20
Hill 95	Phase I				NA	1	2 patients presented serious coronary events after treatment. No positive effect on regional perfusion, contractility, or exercise capacity.
	G‐CSF sc. (16 p.)	10 µg/kg/d for 5 days	52±3	51±3
Joseph 96	Phase I				NA	9	The increase in LVEF was significant in the 4 patients with ischemic cardiomyopathy.
	G‐CSF sc. (6 p.)	5 µg/kg/d for 5 days	19±2 ^j	25±4 ^j

References 33, 51, 61, 89, 90 and 92 included patients with old myocardial infarction and chronic total coronary occlusion, respectively, and patent culprit artery when undergoing stem cell therapy; reference 91 included patients with ischemic cardiomyopathy, symptomatic heart failure, and no options for conventional revascularization; reference 93 included patients with stable coronary artery disease not eligible for or unwilling to undergo bypass surgery (patients with previous myocardial infarction or systolic dysfunction were excluded); reference 94 and 95 included patients with coronary artery disease, severe stable angina, and reproducible ischemia on stress test, not suitable for conventional revascularization and/or unwilling to undergo bypass surgery, respectively; finally, reference 96 included patients with advanced chronic heart failure.

Values are mean±standard deviation or range. LVEF = left ventricular ejection fraction; F‐U = follow‐up; BMMC = bone marrow mononuclear cells; BMNC = bone marrow‐nucleated cells; G‐CSF = granulocyte colony‐stimulating factor; GM‐CSF = granulocyte‐macrophage colony‐stimulating factor; NR control = nonrandomized control group; R control = randomized control group; CPC = circulating progenitor cells; MSC = mesenchymal stem cells; OMI = old myocardial infarction; sc. = subcutaneous administration; i.c. = intracoronary administration; p. = patients; NA = not available.

^a All studies were open label unless stated otherwise.

^b LVEF data showed in the table were calculated or measured by left ventricular angiography 90, 92, 93, 2‐dimensional echocardiography 89, 91, 96, radionuclide ventriculography 51, cardiac magnetic resonance 33, 61, 95, and gated‐SPECT 94.

^c Timing indicates the number of months from myocardial infarction to the administration of the treatment.

^d F‐U indicates the number of months from treatment administration to follow‐up imaging assessment.

^e Including mesenchymal stem cells (66±19%) and endothelial progenitor cells (28±11%).

^f with preconditioning of treated areas by means of induction of previous short ischemia.

^g CPC were intracoronarily infused after recanalization of coronaries with chronic coronary occlusion for 8±3 months.

^h Cell transplantation was carried out immediately after G‐CSF completion.

ⁱ Cell‐therapy trial aimed at assessing the effect of GM‐CSF treatment on angiogenesis, patients with previous myocardial infarction or systolic dysfunction were excluded.

^j LVEF data from the 4 patients with ischemic cardiomyopathy.

Figure 2 The principal differences between cardiovascular stem cell therapy clinical studies have been the type of patients studied, the type of cell used, the delivery approach, and the method of measurement. Three different clinical scenarios have been investigated to date: acute myocardial infarction, myocardial ischemia without revascularization possibilities, and ischemic cardiomyopathy with a depressed contractile function. The type of stem cell used could be classified in two groups: those bone marrow‐derived and myoblast. In conditions where reperfusion is the primary objective or chronic ischemia prevails, the angiogenic potential of bone marrow‐derived may be of high priority. In contrast, in patients where additional restoration of contractile function is the clinical goal, delivering cells with contractile potential like myoblast seems a more reasonable approach. The routes of stem cell delivery experimented have been percutaneous intracoronary infusion, transendocardial delivery through a left ventricle catheter, transvenous catheter approach, transepicardial delivery during surgery, and pharmacological mobilization with stimulating factors. Finally, to assess the efficacy of stem cell therapy, different imaging techniques have been used either to measure perfusion or left ventricular function.

Among the trials performed in patients with acute myocardial infarction, all cases have been performed with the culprit artery patent. Most have been phase I studies analyzing feasibility and safety. The delivery approaches used have essentially been two: the intracoronary injection and the mobilization of stem cells using G‐CSF. The types of stem cells that have been used are bone marrow‐derived. Regarding the design of the trials, most studies have not been randomized, though randomized trials involving a considerable number of patients exist, and although general controversial results have been found, some benefits in ejection fraction improvement or remodeling related parameters have been observed. Recently, the largest study of cardiac cell therapy to date administering hematopoietic progenitors intracoronarily after acute myocardial infarction has been published 26. The REPAIR‐AMI trial, designed as a randomized, double‐blinded study, including 204 patients from 17 European centers, showed a significant increase in the left ventricular ejection fraction (5.5±0.7% versus 3±0.7%; P = 0.014) among patients treated with cells compared to those given placebo. However, other contemporary studies, although reinforcing the message that bone marrow stem cell infusion is not only feasible but also safe, failed to show significant improvement in left ventricular function 27. Importantly, technical differences in the characteristics or handling of the infused stem cells might explain the different outcomes observed 28.

These conflicting results are similar to those found in patients receiving G‐CSF. Four studies published within the last year did not find any differences in the left ventricular function among patients treated with the factor or placebo 29–32. Possible causes to argue these differences could be absence of homing signals, timing of drug administration (in the study which showed a benefit, G‐CSF was administered immediately after myocardial infarction, while in others administration was delayed for 3–5 days after revascularization), subpopulations of stem cells mobilized not effective, and objectives of the study not well established.

Simultaneously to the studies performed in patients with acute myocardial infarction, some groups have tried to avoid established left ventricular remodeling as it happens in the chronic ischemic cardiomyopathy, or to improve myocardial perfusion as it happens in patients with chronic ischemia and no revascularization options. In this chronic clinical scenario the variability of stem cell and delivery approaches used is higher, but in general direct intramyocardial injection (Table II) or a transvascular approach (Table III) have been used. The first trials were done using myoblasts and bone marrow‐derived progenitors directly injected in the infarcted myocardium during CABG surgery. Other studies have evaluated the transendocardial delivery safety also using myoblasts or bone marrow progenitors, and recently experiences have been reported on infusing progenitor cells intracoronarily after opening the coronary artery chronically occluded and with myoblasts transvenously (Tables II and III). However, only one double‐blinded, randomized clinical study in this setting has been reported to date 33. But despite the positive results observed in this study, it is important to notice that no definitive conclusions can be reached due to study design limitations. Thus, we will have to wait until randomized, double‐blinded ongoing studies are fully reported.

Safety

When a new therapeutic strategy gets started, safety becomes the main goal of every single investigation testing that treatment. Undoubtedly, that long‐term stem cell therapy is safe is far from being proved. Therefore, current ongoing randomized mechanistic intermediate‐size trials on stem cell regenerative therapy must focus not only on remodeling and perfusion but also on safety. A shared database with all complications of stem cell treatment in humans should be a priority of all investigators involved in this field. Trials undertaken so far shed light on several complications which cast doubts on regenerative therapy safety in humans: arrhythmias, restenosis and atherosclerosis progression, and nontarget organ seeding.

Arrhythmias have been mainly reported in clinical trials using myoblasts for cardiac repair. It is admitted that myoblasts do not become coupled with the surrounding myocytes providing a putative arrhythmogenic substrate 9, 10. However, it is not clear that these in vitro observations are translated to the clinical setting. Although the first study by Menache observed a high incidence of ventricular arrhythmias in patients with severely depressed ejection fraction 8, no arrhythmias have been reported when myoblasts were cultured without bovine serum 34. Furthermore, in the MAGIC trial the incidence of arrythmias was similar in the active group compared with the placebo group (personal communication). Thus far, this adverse effect could not be exclusive of myoblast. In our ongoing experience 35, we have intracoronarily transplanted autologous bone marrow mononuclear cells in 72 patients following ST‐segment elevation acute myocardial infarction. Four patients showed delayed episodes of ventricular arrhythmias, and three were implanted with an internal defibrillator. Recently, Bartunek et al., intracoronarily delivering CD133+ in patients with acute myocardial infarction, have reported two patients presenting ventricular tachycardia 36. Such adverse events required us to be extremely cautious until future larger controlled trials assist us in identifying the risk attributable to cell replacement therapy and to the arrhythmogenic substrate.

A small clinical trial with ten patients showed a high rate of in‐stent restenosis after G‐CSF administration 37. Whether this finding is related to a wrong timing of G‐CSF injection 4 days before stent implantation remains a matter of conjecture. In this regard, further randomized trials of G‐CSF in patients with acute myocardial infarction treated with primary angioplasty with stent have showed a similar binary restenosis rate in both the active and the control groups 29–32, 38, 39. Regarding bone marrow stem cells, a low restenosis rate has been published in initial studies with angiographic follow‐up 27, 30. A recent nonrandomized matched trial in acute myocardial infarction raises concerns on this issue, though. After intracoronary injection of CD133+ cells, despite a significant improvement in remodeling and perfusion parameters, unexpected rates of 37% in‐stent restenosis and 11% reocclusion were found 36. The use of bare metal stents and CD133+ cells which carry a high angiogenic potential may in part account for these apparent discrepancies.

A rather high incidence of de novo lesions at nonstented arteries has been found after stem cell transplantation 15, 36. The capability of stem cells on angiographically smooth arteries to induce or accelerate atherosclerosis cannot be neglected. Only randomized trials using angiographic and intravascular ultrasound end points will provide information on this topic.

Molecular imaging with radioactively labeled stem cells in humans showed that the percentage cells homing to the heart is quite low with other organs (kidney, liver, and spleen) being the receptors of the vast majority of progenitors 21. The clinical consequences, if any, of this unintended nontarget organ homing are not known.

Stem cell therapy for cardiac repair into perspective

Much excitement has surrounded recent breakthroughs in the field of adult stem cell research and their implications for cell therapy in patients with acute and chronic ischemic heart disease. In the last 5 years numerous clinical studies have been performed in order to evaluate the feasibility and safety of the different stem cell techniques in the acute or chronic cardiovascular setting, with most studies being small‐size and nonrandomized.

Available clinical data show that this therapy is a feasible and safe procedure (with unresolved concerns on arrythmias and restenosis) that seems to diminish significantly the extent of left ventricular remodeling, promote recovery of both global and segmental cardiac function, and improve myocardial perfusion. At this point experts advocate to no longer perform studies involving small numbers of patients, but rather to conduct intermediate‐size, randomized controlled trials to establish the effects of stem cell therapy using surrogate end points, not forgetting safety concerns 40.

These trials should be compatible with simultaneous mechanistic and basic research aimed at addressing some crucial questions still unsettled: How do adult stem cells contribute to cardiac repair? At least six different mechanisms might be proposed: transdifferentiation, de‐differentiation, transdetermination, cell fusion, true pluripotent stem cell behavior, and the production of trophic factors. It is likely that our current knowledge of cell markers is inadequate to define cell populations accurately, so it is possible that cells with true multipotency may have contaminated experiments previously reported as examples of transdifferentiation. What will be the ideal source of adult stem cells? Bone marrow stem cells would seem simple, cheap, and apparently widely applicable. What is the best route of administration? When should cell transplantation be performed? Experimental studies suggest that stem cell transplantation is most successful after the inflammatory reaction is resolved but before scar expansion. How many times should cell transplantation be performed? Several clinical studies have demonstrated that stem cells can improve cardiac function in patients with old myocardial infarction, so successive transplantations may improve the benefit of this therapy.

We are optimistic that stem cell transplantation is likely to be of future benefit and that new tools, such as magnetic resonance, will help to refine patient selection and lead to a better assessment of results. In this sense, as pointed out by others 41, there is no doubt that the success of stem cell therapy will rest on its ability to show clinical efficacy. Thus, the question is whether it is time to perform large‐scale, randomized clinical trials, involving more than a thousand patients, to establish the impact on patient outcomes. In our point of view, that time has not yet come, considering that most studies have included patients with good prognosis, out of the hyperacute phase, and considering the variability in the sources of cells used and the number and quality of cells obtained. Therefore, too much hurry and neutral results could irreversibly damage this promising field of investigation that is currently far from its optimal application method.