|Home | About | Journals | Submit | Contact Us | Français|
Stem cells have been touted as the holy grail of medical therapy with promises to regenerate cardiac tissue, but it appears the jury is still out on this novel therapy. Using advanced imaging technology, scientists have discovered that these cells do not survive nor engraft long-term. In addition, only marginal benefit has been observed in large animal studies and human trials. However, all is not lost. Further application of advanced imaging technology will help scientists unravel the mysteries of stem cell therapy and address the clinical hurdles facing its routine implementation. In this review, we will discuss how advanced imaging technology will help investigators better define the optimal delivery method, improve survival and engraftment, and evaluate efficacy and safety. Insights gained from this review may direct the development of future preclinical investigations and clinical trials.
Heart failure remains the leading cause of death worldwide, accounting for 1 in 9 deaths in the United States in 2007.1 Unfortunately, most current therapies do not address the issue of cardiomyocyte loss and serve only to delay progression or alleviate symptoms until potential orthotopic heart transplantation. Stem cell regenerative therapy has emerged as a promising alternative to improve heart function and prevent the development of end stage heart failure. New advances in stem cell technology have led to an explosion of preclinical and clinical studies using various delivery protocols and cell types, each touting their clinical advantages. Although results from early animal studies and pilot human trials were encouraging, findings from larger randomized studies have been underwhelming. Two recent meta-analyses of 18 and 10 trials showed only marginal benefit,2, 3 and two large-scale randomized Phase II trials using bone marrow-derived cells (BMCs) demonstrated conflicting results.4, 5 The reasons for these puzzling findings remain unclear, but may be further explored by applying imaging technology to guide stem cell therapy. In this review, we discuss how recent advances in imaging technology can help address the many challenges facing the clinical translation of stem cell therapy (Figure 1).
Advanced imaging technology can provide anatomical and functional assessment (e.g., conventional anatomical and functional imaging) as well as visualization of biological processes at the cellular and molecular level (e.g., molecular imaging). Both disciplines utilize various imaging modalities such as radionuclide, optical, magnetic resonance, computed tomographic, and ultrasound technology. Selecting the most effective imaging strategy requires a determination of whether the imaging system can meet the necessary requirements for spatial and temporal resolution, sensitivity, and penetration depth for visualization of the imaging target. Table 1 details the advantages and limitations of different imaging modalities and can serve as a guide for selecting the most appropriate strategy for the chosen indication.
While conventional anatomical and functional imaging can provide an overall road map for cell delivery as well as an assessment of the effects of stem cell therapy, molecular imaging can be used to track stem cells in vivo and to study their potential mechanistic benefits. In molecular imaging, imaging probes are used to target the biological process of interest. These imaging probes consist of a carrier (i.e., a cell, nanoparticle, and microbubble), which structurally binds a ligand designed to recognize the molecular target and a signal element to generate a detectable signal (Figure 2). The ideal imaging probe should have the following important properties: 1) high imaging specificity for tracking the desired biological process, 2) high imaging sensitivity for detection by available imaging modalities, 3) minimal cellular toxicity, and 4) minimal systemic toxicity.
In general, there are two main labeling approaches, each with its unique advantages and disadvantages: 1) direct labeling with radionuclides or iron nanoparticles and 2) reporter gene/probe labeling. Using the direct labeling approach, contrast agents (e.g., signal elements) either bind to cell surface proteins or are transported into the target cell by diffusion, endocytosis, or active transport (e.g., radiolabeled indium oxine and superparamagnetic iron oxide particles) (Figure 3A). In contrast, reporter gene/probe labeling requires cell transfection or transduction with a reporter gene that produces specific proteins (i.e., membrane transport, surface receptor, and intracellular storage proteins as well as intracellular enzymes) that can take up exogenously administered contrast agents. By far the most widely used reporter genes are firefly luciferase (Luc) and herpes simplex virus thymidine kinase (HSV-tk) and their mutants. After delivery of their respective substrates, these enzymes catalyze a chemical reaction that produces a detectable signal (Figure 3B).
The major advantage of reporter gene/probe labeling, especially for in vivo cell tracking, is that cells must be viable with intact protein synthesis machinery in order to produce a detectable signal. In contrast, the signal produced by direct labeling with radioisotopes can be diluted by cell division or dissipate after radioactive decay and/or may persist despite cell death due to the engulfment of dead cells by macrophages.6 Iron labeling by MRI, for example, can remain in the injected site long after cell death, providing erroneous information on the long-term fate of cells despite its superiority in cellular localization.6 Reporter gene/probe imaging is thus better suited for in vivo monitoring of cell viability. In one of the first clinical applications of reporter imaging, the positron emission tomography (PET) reporter probe HSV-tk was used to track and monitor “suicide gene” therapy for gliomas and hepatocellular carcinomas.7, 8 More recently, Yaghoubi et al demonstrated that reporter gene imaging could track the fate of exogenously administered, genetically modified, and therapeutic cytolytic T cells in patients with glioblastoma.9 However, widespread application has been slowed by safety concerns, such as the potential risk of immunogenicity and tumorgenicity caused by random reporter gene integration, as well as limited sensitivity due to reporter gene/probe slicing.10
Both conventional and molecular imaging may help determine the best strategy for stem cell delivery. A lack of an optimized and standardized protocol for safe and effective stem cell delivery is a potential reason for the inconsistent results from previous trials. In preclinical and clinical studies, cells have been delivered via intravenous, intracoronary, or intramyocardial routes.11 Cells have also been administered as early as minutes and as late as a few months post infarction in acute and chronic ischemic models, respectively. Unfortunately, these significant protocol variations have impeded the accurate interpretation of preclinical and clinical trial results, as it is unclear whether the limitation lies with the regenerative capacity of stem cell therapy or with the techniques of delivery.
Stem cells have been delivered via intravenous, intracoronary, and direct intramyocardial routes. Small animal preclinical studies have been restricted to intravenous or intramyocardial delivery due to the small size of the murine and rat coronary arteries. Conversely, the intracoronary approach appears to be preferred in large animal and human trials, despite the lack of evidence for its superiority. Only a few studies have directly compared the cellular retention and engraftment of various delivery routes using in vivo imaging. One such study compared the delivery efficiency of intracoronary compared to peripheral intravenous injection of fluorescence and 99mTc labeled autologous BMCs in a swine model of myocardial infarction.12 For intravenous injection, radioactivity was mainly detected in the lungs with cardiac activity at only minimal levels one-hour after injection. For intracoronary injection, 34.8±9.9% of cells were detected in the heart one-hour post injection, but the number of cells declined precipitously to 6.0±1.7% at 24 hours. Cell detection, however, may be limited by the tracer’s short half-life.
Although intravenous or intracoronary injection is relatively easier and safer, direct intramyocardial injection bypasses the need for cell homing to the injured myocardium, which can be challenging. For example, in a previous clinical trial, only 1.3-2.6% of fluorine-18 fluorodeoxyglucose (18F-FDG) labeled cells were detected in the infarcted myocardium 50 to 75 minutes after intracoronary injection.13 On the other hand, a higher number of cells have been found in the myocardium after direct myocardial injection, but the potential success of this technique may be less predictable.14 To address these limitations, a recent study has found that seeding cells within a fibrin porous biomatrix improved myocardial cell retention in an immunodeficient mouse model of left ventricular (LV) remodeling.15 Taken together, these findings suggest that intramyocardial injection with adjuvant agents to improve cell retention is the preferred delivery method. Greater enthusiasm for this approach may be generated with imaging guidance to accurately delineate the peri-infarct area and to safely deliver cells into the myocardium.
One approach to ensure the safety and consistency of intramyocardial delivery is to use image guided delivery. To derive the most benefit from direct intramyocardial injection, cells need to be physically placed near the target area of injury, and ideally in the border zone between viable and non-viable tissue, but not directly into the infarct. In this location, cells can provide the most benefit and have sufficient access to nutrients from the blood supply. Because the ischemic microenvironment may be less hostile in the peri-infarct area than in the infarct zone, cells may also have a relatively higher chance of survival and engraftment in the peri-infarct area.
Recent advances in imaging technology have enabled image-based navigation of stem cell injection. The best established technique for three-dimensional (3D) intra cardiac navigation is intraventricular electromechanical mapping, which offers superior spatial orientation compared to x-ray fluoroscopy alone. Electromechanical mapping records point-by-point measures of electrophysiologic and motion data, enabling the formation of an electroanatomical map to define the peri-infarct area.16 Although it has been successfully applied in a number of Phase I clinical studies including an ongoing multi-center trial of intramyocardial delivery of bone marrow mononuclear cells,17 the accuracy of electromechanical mapping has been questioned due to overlap of electrical and mechanical data between normal and abnormal myocardium as well as positional artifacts such as heart motion, subject movement, and catheter induced bundle branch block.16 Other disadvantages include a limited access to remote areas, radiation exposure, and lengthy procedure time that can take up to 40 minutes or longer for higher resolution imaging, as more points need to be mapped.16
More recent advances in imaging technology have facilitated the development of two alternative techniques that may be used for defining the peri-infarct area at the time of cell delivery. The first is the emergence of the C-arm computed tomography (CT), which enables multiple serial ECG-triggered rotational acquisitions during slow intravenous contrast injection to achieve a 3D data cube with multi-planar reconstruction of the heart (Figure 1).18 Although this technique has been used to localize areas of injured myocardium induced by radiofrequency ablation,19 further study is needed to determine if the C-arm CT can provide delayed enhancement images of the infarcted myocardium similar to standard CT to enable imaging-based guidance of cell delivery.
The second alternative is the development of software that facilitates co-registration of CT or magnetic resonance imaging (MRI) delayed enhancement images with x-ray fluoroscopy.20, 21 The latter has the advantage of less radiation exposure, greater spatial resolution, and more validated protocols for viability imaging. Tomkowiak et al recently built customized software that has been shown to accurately register and overlay images acquired by MRI with those acquired from x-ray fluoroscopy by aligning discrete anatomical landmarks in a swine myocardial infarction model (Figure 1).21 Further study is required to demonstrate whether this technique can be applied to intramyocardial stem cell injections into the peri-infarct area.
A potential reason for the marginal and inconsistent benefit in cardiac function noted in clinical trials is the variability in cell survival, engraftment, and differentiation. It is likely that functional improvement will require the presence of an adequate number of cells to differentiate into cardiac myocytes, recruit endogenous stem cells, and release paracrine factors to enhance the function of surviving cardiomycytes. Molecular imaging has enabled in vivo monitoring of stem cell fate and may help investigators identify important strategies to improve cell survival, proliferation, engraftment, and differentiation.
Unlike MRI, which can provide only qualitative information on cell location and engraftment, reporter gene-based bioluminescence imaging (BLI) and nuclear imaging can quantify cell survival, proliferation, and engraftment. BLI, however, is limited to small animals due to poor optical transmission. An attractive alternative is PET imaging of 18F-FDG labeled cells, which allows quantification of engraftment in vivo in small animal and large animal models as well as humans.13, 22 However, the half-life of 18F is only 110 minutes and, therefore, only acute evaluation of biodistribution immediately after transplantation is feasible. For serial monitoring, single positron emission computed tomography (SPECT) imaging of 111In-labeled cells, which has a half-life of 2.8 days, can be used to follow subjects two weeks post cell delivery.23 Perhaps the best long-term monitoring strategy for large animals and humans is PET reporter gene imaging with HSV-tk. Because the reporter gene is incorporated into the cell genome, enabling continuous transcription of the reporter enzyme if the cell remains viable, serial imaging can be performed after initial and repeated administrations of the radiolabeled substrate. An attractive alternative to the more commonly used HSV-tk is the sodium-iodide symporter (NIS), which promotes in vivo cellular uptake of 99mTc or 124I for cell tracking using SPECT or PET, respectively. The advantage of NIS is that it is an endogenous mammalian gene that is potentially less immunogenic than HSV-tk. However, the feasibility of this technique has only been demonstrated in small animals.24
Using these various in vivo imaging techniques, studies have revealed that only a limited number of cells engraft and most cells die shortly after transplantation. Specifically, cell tracking studies have found that myocardial engraftment is less than 10% within 48 hours irrespective of cell type, the number of cells implanted, and delivery route.25, 26 The majority of cells are not found in the myocardium, but are either trapped in the pulmonary or microvasculature27, 28 or localize to remote organs.14, 29 Cell homing and retention in the infarcted myocardium, however, does occur. In a mouse model, Sheikh et al demonstrated that bone marrow mononuclear cells from a male donor injected intravenously preferentially home in on and are retained in the myocardium in female mice with ischemia-reperfusion injury compared to sham mice.30 The authors used in vivo cell tracking by BLI and ex vivo quantitative real time polymerase chain reaction analysis for the male Sry gene four weeks post injection. Interestingly, in a dog model of myocardial infarction, Kraitchman et al showed that intravenous injection of mesenchymal stem cells (MSCs) redistribute from the lung to infarcted myocardium,31 one-day post injection, further supporting that stem cells home to the injured tissue. Nevertheless, at present, the degree of cell homing and retention appears inadequate.
Even more disappointing is the fact that most cells die shortly after transplantation.6, 32-36 After delivery of human embryonic stem cell (ESC)-derived cardiomyocytes in an immunodeficient mouse model of myocardial infarction, in vivo BLI demonstrated that 90% of cells died within 3 weeks of delivery.34 Other investigators have also demonstrated failure of long-term survival (>8 weeks post delivery) for additional cell types, including MSCs,33, 36 skeletal myobalsts,33 cardiac resident stem cells,35 bone marrow mononuclear cells,30 adipose stromal cells,36 and ESC-derived endothelial cells32 in small animal models (Figure 4). Similarly, in a porcine model of myocardial infarction, Gyöngyösi et al could only detect faint focal activity in the myocardium 7 days post injection of porcine MSCs transfected with a PET reporter gene (Figure 5).37 Failure of cell detection beyond 7 days in large animal models is likely due to limited resolution and sensitivity of PET for long-term monitoring. As the cell number declines over time, PET may be unable to detect the generated signal, especially given that only a fraction of cells are labeled by reporter genes due to variable viral transduction efficiency as well as reporter gene/probe silencing.26 Taken together, these findings emphasize the need to understand the reasons behind poor cell retention and survival in order to identify methods for their improvement.
In addition to poor survival and engraftment, directed differentiation of pluripotent stem cells into functional cardiomyocytes remains challenging. Cardiac differentiation appears to be enhanced by prior in vitro induction, which was confirmed by a study using BLI and stem cells expressing fluorescent and Luc genes under control of the cardiac sodium-calcium exchanger-1 (Nxc-1) promoter.38 Cells that showed markedly enhanced Luc expression upon induction of differentiation in vitro continued to express Luc for 2-4 weeks post transplantation in vivo. This finding has been supported by the emergence of several differentiation protocols for cardiomyocyte generation from pluripotent stem cells, using various growth factors, chemical, and physical stimuli.39 Despite these efforts, cardiomyocyte differentiation in vitro is largely uncontrolled and inefficient, with a success rate of <25%, under most protocols,40 highlighting the need for better understanding of the mechanisms regulating differentiation. Interestingly, one study has elucidated the role of STAT3 activity in ESC differentiation using BLI and a custom-designed STAT3 reporter construct driving a fluorescent protein and Luc.41 Subject to further investigation, in vivo imaging of cardiac differentiation has the potential to provide valuable insight into the control of stem cell differentiation.
By providing greater understanding of the mechanism behind poor engraftment and inefficient differentiation, in vivo imaging can guide the development of techniques to address these limitations. In vivo imaging has evaluated the optimal timing for stem cell delivery (acute vs. sub acute infarction),42 determined the spatiotemporal kinetics of BMC homing,30 compared delivery of various stem cell types,33 and assessed the potential role of pro-angiogenic or pro-survival agents to meet these limitations.43, 44 For example, Swijnenburg et al used in vivo cell tracking to determine whether timing of cell delivery would affect cell viability, but found no significant difference in cell survival or cardiac function when cells were delivered immediately (acute) versus 7 days (sub acute) after myocardial infarction (Figure 4C).42 In another study, acute donor cell death was seen in all cell types when comparing engraftment rate of bone marrow mononuclear cells, MSCs, skeletal myoblasts, and fibroblasts (Figure 4B).33 More recently, Hu et al showed that the addition of a microRNA prosurvival cocktail (miR-21, miR-24, and miR-221) was associated with enhanced survival of cardiac progenitor cells by BLI and improved cardiac function by echocardiography.44
Using BLI, investigators have also determined the kinetics of donor cell rejection across different immunological barriers.45-47 Swijnenburg et al demonstrated that survival of transplanted human ESCs was significantly limited in immunocompetent as opposed to immunodeficient mice, but could be mitigated by administration of immunosuppressive agents.45 Because these traditional regimens have shown only marginal improvement in survival, a recent study demonstrated that brief treatment with three co-stimulatory receptor blocking agents (i.e., cytotoxic T-lymphocyte-associated antigen 4 (CTL4)-Ig, anti-CD40 ligand, and anti-lymphocyte function-associated antigen 1) induced long-term engraftment of mouse ESCs, human ESCs, mouse iPSCs, human iPSCs, and more differentiated ESC- and iPSC-derivatives.47 As an alternative to adjuvant immunosuppressive therapy, multiple investigators are exploring whether the administration of autologous iPSCs, which are reprogrammed from somatic cells, may be more immunoprivileged. The feasibility of transplantation of autologous iPSCs in large animal models was recently reported using combined PET and MRI (Figure 6).48 Whether iPSCs are immune tolerated, however, has been recently challenged by a study that demonstrated that iPSCs could also induce a T-cell dependent immune response in syngeneic recipients, despite their autologous origins.46 Overall, these studies underscore the importance of in vivo imaging in developing strategies to enable the clinical translation of stem cell therapy.
It was previously believed that the principal mechanism underlying the benefit of stem cell therapy was the regeneration of functional endothelial cells (ECs), vascular smooth muscle cells (VSMCs), and cardiomyoctyes. Early studies in murine models of myocardial infarction have shown that cardiac stem cells appeared to integrate into the surrounding native myocardium and regenerate damaged tissue.49 With the advent of in vivo imaging, it has become evident that <10% of cells are retained in the myocardium 24 hours post injection and most injected cells do not survive, engraft, or proliferate beyond 8 weeks.25, 26 Because of the number of newly regenerated cells is too low to explain the significant functional improvement observed in most preclinical and clinical studies, it has been suggested that the functional benefit of therapy is related to secretion of soluble factors that protect the heart, induce neovascularization, and attenuate pathological remodeling (i.e., paracrine effect).49 Although the exact mechanism associated with improvement in LV function remains unclear, molecular imaging may help elucidate some of these unknowns.
A potential application of in vivo imaging is to provide confirmation that stem cells release cytoprotective molecules that increase cardiomyocyte survival (i.e., factors that reduce apoptosis and necrosis). Data in animal models suggest that even very low levels of apoptosis (23 myocytes per 105 cardiac nuclei) could result in progressive lethal dilated cardiomyopathy that may be prevented by the administration of stem cell therapy.50 The feasibility of in vivo imaging of cell death has been demonstrated using 99mTc labeled annexin albeit with difficulty differentiating necrosis from apoptosis.51 These limitations may be addressed using a combined annexin-labeled magnetofluorescent nanoparticle (AnxCLIO-cy5.5) with gadolinium-DTPA-NBD, to detect apoptosis and necrosis, respectively.52 Incorporating methods to image stem cell death may help develop future therapies to improve stem cell survival.
Another important application of in vivo imaging is to verify that stem cells promote angiogenesis and arteriogenesis via release of paracrine factors, given that only a small number of vessels contain donor cells.49 In animal models of hindlimb ischemia and myocardial infarction, stem cell administration was associated with increased expression of pro-angiogenic factors,53, 54 resulting in an increase in capillary density and collateral development, that can occur prior to any significant improvement in blood flow.55 In support of these findings, in vitro imaging using time elapse microscopy has shown that co-culture of ESCs with ECs results in EC migration, proliferation, lumen formation, and anastamosis to existing vasculature.56 The feasibility of in vivo imaging of angiogenesis in large animals and humans has already been demonstrated using radionuclide probes targeting αvβ3 integrin (i.e., 18F-Galacto-RGD and 123I-Gluco-RGD),57, 58 which are receptors that mediate endothelial cell migration, proliferation, and survival. Although it has yet to be applied directly to stem cell imaging, this technique has been used to demonstrate neovascularization after treatment with vascular endothelial growth factor in a porcine model of myocardial infarction, paving the way for its application to assess the benefits of stem cell therapy (Figure 7).58 It will be important to apply these imaging techniques in future studies to correlate the degree of angiogenesis with improvement in LV function.
In vivo imaging may also help us better understand how stem cell therapy attenuates pathological remodeling. Stem cell administration has been found to decrease fibrosis, resulting in improvement in LV dilatation and systolic and diastolic function.59 These positive effects on post-infarct remodeling may be mediated by release of molecules that limit local inflammation (e.g., monocyte chemoattractant protein) and factors that modulate proliferation of fibroblasts and synthesis of collagen and extracellular matrix (i.e., metalloproteinases, tissue inhibitor of metalloproteinase, transforming growth factor, serine proteases, and serine protease inhibitors).60 Imaging of leukocyte trafficking, protease activation, collagen deposition, and myocardial fibrosis in animal models of myocardial infarction has already been achieved using various imaging techniques, including fluorescence molecular tomography, MRI, and radionuclide imaging.60 Future studies evaluating these processes before and after stem cell therapy will improve our understanding of how stem cell therapy minimizes scar formation and prevents heart failure progression.
In addition to reducing scar formation, stem cell administration has been shown to reverse abnormal myocardial energetics that has been associated with heart failure. Progressive LV dilatation results in increased LV wall stress, myocyte overstretching, and a shift toward expression of fetal contractile proteins, resulting in imbalance in ATP delivery and demand.61 Administration of BMCs has resulted in improvement of metabolism in both small and large animal models, as measured by nuclear magnetic resonance spectroscopy.62, 63 The development of an ultrafast magnetization saturation transfer method may facilitate examination ATP kinetics in humans in future clinical trials.64
Over the last decade, significant resources have been invested in the development of stem cell therapy for cardiac regeneration, yet results from large animal and human studies based on analysis of global left ventricular ejection fraction (LVEF) have been underwhelming. In a recent meta-analyses of 52 published large animal preclinical studies (n=888 animals, median follow-up=4 weeks with range 2-17 weeks) using various types of cardiac stem cells, treated animals showed improvement in LVEF by 7.5%, resulting from a significant decrease in end systolic volume.65 Important predictors of improvement in LVEF include the use of MSCs, left anterior descending artery infarction, chronic occlusion models, higher number of cells (≥107), and cell injection at least 1 week after myocardial infarction. Animal type and route of delivery were not predictive. Similarly, two meta-analyses of clinical studies (n=10 studies with 698 patients and n=18 studies with 999 patients) using BMCs showed an increase in LVEF of 3-4% and a reduction in end systolic volume of 4-6 ml with median follow-up of 6 months (3 to 18 months), measured predominantly by echocardiography and MRI.2, 3 As shown in Table 2, the primary endpoint of most clinical studies has been the evaluation of LV size and global LVEF. Perhaps incorporating additional functional and anatomical measures, such as regional systolic function, diastolic function, perfusion, infarct size and viability, will further elucidate the benefits of stem cell therapy.
The majority of preclinical and clinical studies have measured changes in LV size and function pre- and post-treatment to evaluate the effectiveness of cardiac regenerative therapy. The accurate and precise measurement of these parameters is critical given the relatively small changes observed after stem cell therapy. Unfortunately, M-mode (i.e., fractional shortening) and two-dimensional echocardiography (i.e., Simpson’s rule) make geometrical assumptions about the left ventricle that can lead to inaccurate measurements, especially in the presence of regional wall motion abnormalities. Poor inter- and intra-observer variability also limit the use of these techniques for serial imaging. These limitations can be addressed by the use of 3D echocardiography and cardiac MRI, which provide improved accuracy and reproducibility.
An alternative approach, and perhaps a more sensitive marker, is to assess regional contractile function by measuring strain and strain rate by echocardiography or MRI. The strain and strain rate reflect the absolute deformation and the speed of deformation (shortening and thickening) of the myocardium from an applied force in a specific myocardial segment and layer (i.e., endocardial and epicardial walls), respectively. This technique subtracts motion due to the effects of neighboring segments (e.g., tethering) that can mask pathological deformation and impart abnormal motion to normal segments. Abnormal strain and strain rate imaging, for example, have been found in patients with diabetes, amyloidosis, and doxorubicin-induced cardiac injury, who have normal LVEF.66 Similarly, a recent report in 67 ST-elevation myocardial infarction patients randomized to intracoronary infusion with bone marrow progenitor cells or placebo showed significantly greater improvement in regional strain rate in the treatment group with no significant differences in LVEF (Figure 8).67 Further application of this technique in preclinical and clinical studies may reveal a more impressive contribution of stem cell therapy for improving contractile function.
In addition to systolic dysfunction, an abnormal LV filling pattern is common in patients with ischemic heart disease and is a marker of poor prognosis. Moreover, patients who have improvement in LV filling after treatment have fewer symptoms and better survival.68 Despite this evidence, the evaluation of diastolic function has been performed in only a handful of clinical studies, unlike in animal studies where invasive PV loop measurements of LV relaxation are mainstay. In an initial study, Schaefer et al reported improvement in mitral valve filling (E/A ratio) for up to 18 months in patients after acute myocardial infarction, who were randomized to treatment with intracoronary infusion of bone marrow cell therapy, but the effect lasted less than 5 years.69 Other diastolic function parameters showed no difference. Dieterichsen et al also showed improvement in diastolic function (i.e., E/e’ ratio left atrial volume, and plasma NT-pro-BNP) in patients with chronic ischemic heart disease after repeat intracoronary administration of BMCs at 12-month follow-up.70 Although echocardiographic measures of diastolic function are noninvasive and, therefore, preferred for serial evaluation in patients, they correlate only modestly with invasive pressure volume loop analysis, the gold standard for evaluating LV relaxation. Perhaps further development and validation of strain imaging by echocardiography or MRI, which has the potential to be a more accurate and precise measure of LV relaxation properties, will generate greater interest for the incorporation of noninvasive diastolic function measures in clinical stem cell trials.71, 72
The potential for stem cell induced vasculogenesis has prompted several studies to measure myocardial perfusion. Although the majority of studies appear to show improvement, more in-depth analysis reveals that the data are both limited and conflicting (Table 2). For example, only six studies evaluate both stress and rest phases,73-81 and only two studies incorporate PET data,82, 83 which provide a more quantitative assessment of myocardial perfusion. Results have also been inconsistent. Although SPECT perfusion data at rest have shown a decrease in the number of segments with perfusion defects after treatment, studies incorporating both stress and rest phases only show a decrease in overall ischemia extent, but not in the number of rest perfusion defects. Even more discouraging is that only one of the two studies incorporating PET perfusion showed improvement in overall myocardial blood flow in the treatment compared to the control group. Interestingly, all studies evaluating regional myocardial blood flow by intracoronary Doppler wire showed improvement in coronary flow reserve in the infarct related vessel after treatment.84-90 Performing serial evaluation with intracoronary Doppler wire, however, is not without risk. Taken together, these findings suggest that further evaluation of regional as well as global myocardial perfusion reserve by PET should be performed to confirm that stem cell therapy contributes to detectable levels of vasculogenesis in the myocardium.
Although a variety of methods can be used to assess the effects of stem cell therapy on infarct size and the amount of viable tissue, direct visualization by contrast enhanced MRI, radionuclide imaging (i.e., PET or SPECT), and electromechanical mapping are preferred over indirect approaches that evaluate the extent and severity of LV dysfunction, such as LV angiography, 2D echocardiography, or cine MRI. Because of its precise delineation of scar tissue and the ability to distinguish between subendocardial and transmural infarction, contrast enhanced MRI is considered the most accurate for assessing infarct size. Interestingly, the majority of stem cell studies using contrast enhanced MRI have failed to show a significant change in infarct size,4, 82, 84, 86, 91-93 whereas, other techniques have demonstrated a significant decrease (Table 2).
The reasons for this discrepancy remain unclear because previous studies have provided only limited details on how contrast enhancement images were acquired and analyzed, two critical factors for determining changes in infarct size. One reason may be prior studies failed to control for the time interval between contrast injection and image acquisition, which is known to affect infarct size, especially in acute settings where gadolinium can enhance the area at risk as well as the infarct core.94 Additionally, previous studies have used traditional methods of image analysis, which have poor reproducibility.95 Finally, these studies did not evaluate changes specific to the peri-infarct area, which has been shown to have independent prognostic value beyond LV infarct size, volume, or ejection fraction.96, 97 Perhaps a more standardized and comprehensive approach to image acquisition and analysis will provide clearer evidence that stem cell therapy reduces the size of both the peri-infarct and infarct zones.
To determine viability, 18F-FDG PET, which measures glucose metabolism, is considered the gold standard for its superior sensitivity and specificity compared to other functional measures of viability, including stress echocardiography and SPECT.98 All clinical studies using 18F-FDG PET have consistently shown improvement in the myocardial viability (Table 2).83, 88, 99-101 In contrast to glucose metabolism, improvement in contractile reserve measured by dobutamine echocardiography has not been shown in all studies.75, 102, 103 This is not surprising, because myocardium that is severely dysfunctional due to extensive cellular damage usually does not have intact contractile reserve but may have preserved glucose metabolism.104 Additional studies incorporating both these measures will be important to further delineate the benefits of stem cell therapy in improving myocardial viability.
In addition to establishing the efficacy of stem cell therapies, routine clinical implementation of novel cell based treatment will depend on the successful resolution of important safety concerns such as cell migration, tumorgenicity, and arrhythmogenicity. Particularly challenging is the monitoring of cell biodistribution because administered cells may be essentially indistinguishable from host cells. As discussed previously, short-term in vivo cell monitoring has already been achieved in clinical trials using radionuclide based imaging techniques. Safety concerns, however, continue to impede regulatory approval of iron oxide and reporter genes labeling techniques for longer term monitoring. Also problematic is the detection of the low number of cells that may migrate to non-target organs. This need may be potentially met by a recently approved class of fluorocarbon-based reagents designed to safely and efficiently label cells ex vivo without the use of transfection agents.105
The ability to track stem cells in vivo is also vital for detecting the formation of inappropriate ectopic tissues and for guiding their elimination. This is especially critical for human ESCs and iPSCs, which have the potential to form teratomas (tumors composed of a haphazard array of somatic cell types) and teratocarcinomas (a malignant tumor composed of a teratoma mixed with embryonal carcinoma and/or choriocarcinoma).106 Although multi-modality approaches have been applied to image teratoma formation in vivo using MRI and reporter gene-based optical imaging, the former is limited by sensitivity, and the latter by poor tissue penetration. An alternative approach is to use the PET reporter probe 64Cu-radiolabeled cyclic arginine-glycine-aspartic acid (RGD) peptides to image αvβ3 integrin expression, whose up-regulation is known to play a key role in angiogenesis and metastasis (Figure 1).107 Cao et al recently demonstrated the superiority of this technique for monitoring tumorgenicity after ESC transplantation compared to more commonly used PET tracers (18F-FDG and 18F-FLT), which are less sensitive in detecting teratoma formation.108 Finally, molecular imaging can also monitor suicide gene ablation of teratomas using HSV-tk as both a PET reporter gene and suicide gene, respectively.109
Quantification and localization of cell engraftment will be important to limit arrhythmogenic risk. Although higher levels of cell engraftment may be associated with improved efficacy, focal grafts can create inhomogeneities within the myocardium that predispose the heart to re-entrant arrhythmias. In support of this theory, in vitro studies using skeletal myoblasts, neonatal cardiomyocytes, and MSCs have found that the risk of re-entrant arrhythmias appears to be dose dependent.110 Results from subsequent in vivo Phase I clinical trials, however, have been conflicting with an increase in arrhythmogenic risk associated with skeletal myoblasts,111 whereas, minimal risk has been found with transplantation of MSCs and BMCs.110 However, with so few cells surviving long-term in vivo, to conclude that the arrhythmogenic risk of stem cells other than skeletal myoblasts is low would be premature. Overall, these studies in cell migration, tumorgenicity, and arrhythmogenicity demonstrate that in vivo imaging can help investigators monitor recipients after transplantation.
With parallel developments in conventional and molecular imaging technology, investigators are well positioned to apply and extend current capabilities to facilitate the clinical translation of stem cell therapy. These invaluable tools have already helped researchers study issues related to poor engraftment, identify possible mechanisms of benefit, and characterize potential safety concerns. With this knowledge in hand, investigators can now focus on finding solutions to the challenges facing the clinical implementation of regenerative therapy.
In the next several years, we anticipate that investigators will gravitate toward cell delivery under direct visual guidance using advanced imaging modalities, including MRI and CT. In vivo tracking of cells during and shortly after delivery will enable identification of techniques to improve initial engraftment such as administration of agents that optimize cell adhesion, control hemodynamics, and prevent migration to non-target areas. Longer-term tracking of cells remains a challenge, especially in humans, given the limitations of the direct cell labeling and the risks of insertional mutagenesis or immunogenicity associated with reporter gene labeling. Greater focus on the development of a safe, robust technique to label and follow cell viability, differentiation, and migration is needed prior to routine implementation. In the meantime, currently available techniques will be used to identify the optimal cell type, timing of delivery, and immunosuppressive regimens in animal models to improve cell engraftment and survival. Future studies will also likely apply advanced imaging technology to provide mechanistic insight to determine whether regulation of these processes may augment the benefits of stem cell therapy. Finally, we encourage investigators to provide a more comprehensive evaluation of benefits of stem cell therapy beyond an evaluation of global systolic function. We firmly believe that application of the advanced imaging techniques reviewed here will be instrumental to the successful clinical implementation of this novel therapy.
We thank Amy Morris for preparing the illustrations. Due to space limitations, we are unable to include all of the important papers relevant to induced pluripotent stem cell derivation and application; we apologize to those investigators whom we omitted here.
FUNDING SOURCES This work was supported by the ACC/GE Career Development Award (PKN); NIH HL099117 and NIH EB009689 (JCW).
The authors have declared that no conflict of interest exists.