|Home | About | Journals | Submit | Contact Us | Français|
Arteriosclerotic cardiovascular diseases are among the leading causes of morbidity and mortality worldwide. Therapeutic angiogenesis aims to treat ischemic myocardial and peripheral tissues by delivery of recombinant proteins, genes, or cells to promote neoangiogenesis. Concerns regarding the safety, side effects, and efficacy of protein and gene transfer studies have led to the development of cell-based therapies as alternative approaches to induce vascular regeneration and to improve function of damaged tissue. Cell-based therapies may be improved by the application of imaging technologies that allow investigators to track the location, engraftment, and survival of the administered cell population. The past decade of investigations has produced promising clinical data regarding cell therapy, but design of trials and evaluation of treatments stand to be improved by emerging insight from imaging studies. Here, we provide an overview of pre-clinical and clinical experience using cell-based therapies to promote vascular regeneration in the treatment of peripheral arterial disease. We also review four major imaging modalities and underscore the importance of in vivo analysis of cell fate for a full understanding of functional outcomes.
With increasing prevalence, arteriosclerotic cardiovascular diseases are among the leading causes of morbidity and mortality worldwide . Peripheral arterial disease (PAD) in the lower limbs is a common manifestation of systemic atherosclerosis, and is most severely manifested as critical limb ischemia (CLI), which, in the absence of revascularization, represents a 50 percent risk factor for amputation within one year . Even in the absence of symptoms, patients with PAD have an increased rate of stroke and myocardial infarction, and a 600% higher death rate compared to patients without PAD . Furthermore, comorbid PAD increases the risk conferred by coronary artery disease or cardiovascular disease . The standard therapeutic approach for CLI is surgical bypass or endovascular techniques that aim to restore blood flow to the extremity. Despite a five-year limb salvage rate of over 80%, surgical bypass results in a 5% mortality rate postoperatively, and 30 to 50% of cases have complications. For an additional 40% of patients, limb salvage surgery is not a viable treatment option , leading to major amputations in approximately 30% of all cases. This amounts to ~100,000 amputations annually in Europe and ~120,000 in the US .
The need to develop novel treatment approaches has led to therapeutic angiogenesis, which aims to promote neovascularization and improve perfusion of ischemic tissue by forming new blood vessels from preexisting ones. Pioneered by Dr. Judah Folkman over four decades ago , therapeutic angiogenesis today aims to treat vascular insufficiency by augmenting collateral artery development in ischemic myocardial and hindlimb tissues via delivery of recombinant proteins, gene transfer, or cell therapy to promote release of angiogenic factors . The aim of such therapies is to restore the essential vascular network that underlies successful tissue repair— to lessen symptoms, regenerate damaged tissue, and avoid limb amputation. Though only limited clinical data are available at this point, recombinant protein and gene transfer approaches appear to be safe and therapeutically promising. Nonetheless, concerns remain over both the potential development of pathological angiogenesis as well as “bystander effects” on the kidneys or atheromas due to uptake of the therapeutic factor by cells in the target cells’ milieu . Furthermore, despite encouraging data from early Phase I clinical trials, data from Phase II and III trials in myocardial angiogenesis have not confirmed conclusively that gene therapy is a successful approach . As an alternative, cell-based therapies have been shown to promote vascularization of ischemic tissue, as corroborated by observed functional improvements in both stroke and myocardial infarction models [11, 12, 14]. In 1997, Asahara and colleagues demonstrated that endothelial progenitors cells (EPCs) isolated from human peripheral blood incorporated into tissues undergoing active angiogenesis in a murine ischemia model . Follow-up work showed that bone marrow-derived mononuclear cells (BMMNCs) do contain EPCs as well as cytokines and angiogenic factors that underlie BMMNCs’ ability to augment collateral vessel formation (arteriogenesis) . Preclinical work in mice has demonstrated the ability of BMMNCs to integrate with microvasculature and to produce angiogenic cytokines that increase the density of microvasculature as well as perfusion within ischemic tissue [15, 16].
Small and large animal studies have provided evidence for the ability of cell therapy to restore blood flow to ischemic limbs. Early studies demonstrated incorporation of EPCs into interstitial arteries and capillaries in hindlimb models . In athymic mice with hindlimb ischemia, transplantation of ex vivo-expanded EPCs allowed blood flow recovery, increased collateral density, and produced 60% limb salvage, compared to 7% in the control . Most animal models, however, rely on acute rather than chronic ischemia models. This is a significant limitation, because successful translation of therapeutic approaches relies on the ability to treat degenerative arteriosclerotic disease. A summary of small animal studies [13, 19–31] using cell therapy to treat hindlimb ischemia can be found in Table 1.
Despite this limitation, clinical trials using cell therapies originally tested in acute hindlimb ischemia models have produced promising results in treating PAD. In the first large trial of bone marrow cell therapy to treat hindlimb ischemia, the Therapeutic Angiogenesis by Cell Transplantation (TACT) trial reported that implantation of BMMNCs in the limbs of patients with CLI extended the amputation-free interval compared to patients not receiving angiogenic cell therapy . In the randomized, controlled arm of the study, patients received either BMMNCs or peripheral blood mononuclear cells (PBMNCs) as a control. While BMMNC injection led to decreased pain and improved ankle-brachial index, PBMNC injection had no therapeutic benefit. The authors of the study suggested that this difference is due to the greater contribution of CD34+ hematopoietic stem cells and greater number of immature precursor cells in the BMMNC population than in the PBMNC population . Additional studies have demonstrated that BMMNC therapy improves endothelial-dependent vasodilation . A meta-analysis of 37 clinical trials, pooling data from approximately 700 patients who were not candidates for traditional revascularization interventions, concluded that cell therapy significantly improved ankle-brachial index, transcutaneous oxygen tension, ulcer healing, pain-free walking distance, rest pain, and limb salvage as trial endpoints, but cautioned that examination of the findings of trials currently underway is necessary to confirm the efficacy and safety of cell-based therapies .
Technological advances in non-invasive imaging methodologies offer the ability to better understand the interactions between transplanted cells and host tissue that underlie vascular regeneration. The success of cell-based therapies, both in basic science and clinical settings, relies on the ability to track the location, engraftment, and survival of the exogenous cell population, as well as to assess potentially negative effects on host tissue . Without the ability to longitudinally profile the fate and activity of transplanted cells, it is difficult to attribute the observed functional outcomes in animal studies and human trials to the action of the transplanted cell population. The goal for tracking and imaging cell-based therapy is to integrate and correlate these results with functional improvement in limb perfusion. The past decade of investigations has generated promising clinical data regarding cell therapy, and imaging studies have become an essential tool for improving and evaluating the design of trials and specific treatments. For example, most clinical trials involving cardiac stem cell therapy have relied on one or more imaging modalities as prognostic outcome measures [36–40]. By contrast, clinical translation of cell-based therapy for the treatment of peripheral arterial disease has largely focused on extending the amputation-free interval rather than on correlating this measure with cell fate .
Several systems have emerged that allow longitudinal in vivo cellular imaging and offer clear advantages over traditional microscopy methods. Conventional histopathological and cytological techniques require chemical fixation following invasive removal of tissues, provide only non-quantitative data under non-physiological conditions, and generate minimal information regarding dynamic cellular processes . Furthermore, such techniques preclude repeated data collection from the same subject. The transition from in vitro to in vivo models has focused on the development of small animal, non-invasive, high-resolution models . However, interpretation of these data in light of future clinical applications must meet the challenges of subject size, the volume of tissue to be imaged, the requisite spatial resolution, and the scanning time . An analysis of the currently available modalities can shed light on the challenges ahead in developing clinically relevant techniques to image vascular regeneration. The present review examines the benefits and limitations of four systems for evaluating regeneration induced by cell therapies: radionuclide imaging, magnetic resonance imaging (MRI), bioluminescence imaging (BLI), and fluorescence imaging Fig. (1).
Non-invasive molecular imaging strategies utilize two primary methods. Direct (physical) labeling relies on detection of bound extracellular or intracellular probes, whereas genetic labeling relies on stable or transient transfection of cells to express a receptor, protein, or enzyme, which is then detected by reporter gene imaging. Systems that allow for either genetic or physical label detection include positron emission tomography (PET), single photon emission computed tomography (SPECT), magnetic resonance imaging (MRI), and use of an optical charged coupled device (CCD). Direct labeling, using iron particles or radioactive tracers, has been used primarily for radionuclide imaging and MRI, and to a lesser extent for nanoparticles (for example, in quantum dot imaging) . Genetic labeling has been used primarily for PET [45, 46] and SPECT  reporter gene imaging.
The principal advantage of using a physical label is that it requires minimal manipulation of the cell population. The principal disadvantage is that the label can decouple from the transplanted cell so that label detection may not provide accurate data on cell viability or engraftment . In this respect, genetic techniques are superior because the reporter gene is not detected in non-viable cells. The reporter gene, if stably transfected, can be mitotically passed on to progeny cells. Additionally, placement of the reporter gene downstream of a tissue specific or inducible promoter allows for detection of transgene expression under various experimental conditions. The primary disadvantage of genetic techniques is DNA modification of the cell population, which may lead to adverse cellular effects. Overall, the four systems reviewed here each has its own distinct advantages and disadvantages, differing in spatial and temporal resolution, energy use per image generation (ionizing or deionizing), depth penetration, probe biocompatibility, and probe detection threshold  (highlighted in Table 2).
PET and SPECT provide high-sensitivity information about in vivo cell biodistribution. Radionuclide imaging utilizes either physical or genetic labels to detect gamma rays emitted from the subject. Due to the absence of background signal, radionuclide imaging can provide accurate information from small cell populations. PET technology relies on labeling molecules with positron-emitting isotopes. When a positron is emitted from the nucleus, it interacts with an electron to produce two gamma-rays at 511,000-eV and approximately 180 degrees apart. Commonly-employed direct labeling isotopes for PET are 15O, 13N, 11C, and 18F. By contrast, SPECT relies on gamma-emitting isotopes such as 99mTc, 111In, 123I, and 131I, which produce gamma-rays that are detected by cameras rotating around the subject. PET has a sensitivity in the range of 10−11 to 10−12 mole/L, whereas SPECT has sensitivity in the range of 10−10 to 10−11 mole/L .
For physical labeling, cells are incubated with radionuclide agents, which results in a range of labeling efficiency. For example, the labeling efficiency of stem cells after incubation with 2-[F-18]-fluoro-2-deoxy-D-glucose ([18F]-FDG; PET radiotracer) is reported to range from 40 to 100 percent [48, 49]. The SPECT radiotracer [In-111]oxyquinolone (oxine) (111In), with a half life of nearly three days, has been used clinically for more than two decades to localize infections, because labeled leukocytes home in on the site of infection . Both PET and SPECT provide quantitative data on the level of radiotracer, which reflects the number of labeled cells within the region of interest . More recently, radionuclide imaging has been used in a porcine myocardial infarction model to track the fate of intravenously-delivered mesenchymal stem cells (MSCs), but accumulation of the labeled cells in the lung led the authors to recommend different administration routes . Similarly, PET imaging to monitor homing of [18F]-FDG-labeled BMMNCs showed that intravenous cell delivery did not lead to targeting of the infarcted human myocardium. By contrast, intracoronary delivery of a selected cell population resulted in 1.3 to 2.6% of cells homing in on the myocardium. Approximately 14 to 39% of a further selected, CD 34+-enriched population accumulated in the myocardium following intracoronary delivery . The aforementioned studies highlight the sensitivity with which PET and SPECT direct labeling strategies for both basic and clinical use can identify both transplanted cells and regenerating tissue.
Genetic labeling strategies to track transplanted cells are based on enzymes, transporters, or receptors. Reporter gene labeling approaches include the use of human sodium iodide symporter (hNIS)  or herpes simplex virus type 1-thymidine kinase (HSV1-tk) , and are superior to transporter- or receptor-based techniques, which suffer from signal amplification due to probe accumulation . In a porcine model, MSCs were transduced in vitro with HSV1-tk using a lentiviral  or adenoviral  vector prior to transplantation into the myocardium. PET was used to track the persistence of viable transplanted cells after injection of PET reporter probe (9-(4-[18F]-fluoro-3-hydroxymethylbutyl)-guanine ([18F]-FHBG), and demonstrated a decrease in the number of surviving cells, leading to the conclusion that there was cell death and minimal proliferation following intramyocardial injection . Following the Food and Drug Administration’s (FDA) approval of the PET reporter probe [18F]-FHBG as an Investigational New Drug , a 2009 study reported on its use in a clinical trial. Cytolytic CD8+ T cells engineered to express the interleukin 13 zetakine gene (to target T cells to the tumor sites) and HSV1-tk were expanded ex vivo and autologously delivered to a 57-year old male with grade-IV glioblastoma multiforme. [18F]-FHBG was detected five weeks later at the tumor site, highlighting the efficacy of the PET reporter gene strategy . Finally, dual-isotope SPECT has also been used to follow reporter gene expression in pancreatic islet cell transplantation using 131/123I-FIAU as a probe . Other reports have demonstrated that PET with 123I-FIAU imaging can accurately assess the efficacy of ganciclovir therapy to eliminate NG4TL4 sarcoma cells expressing HSV1-tk .
The sensitivity and safety of PET and SPECT have contributed to their success in basic research and clinical settings. Despite their longstanding use, [18F]-FDG and [111In]-oxine labeling approaches for cell-based therapies have several limitations. These include decoupling of the tracer from the transplanted cell, potential adverse effects on cell viability, as well as the inability to track cell fate long-term due to the short half-lives of PET tracers. While genetic labeling approaches offer valuable information regarding long-term cell fate, the introduction of foreign genetic materials into the genome could have significant deleterious effects on the donor population, and might elicit an immune response from the host should the manipulated cells express foreign antigens.
While less sensitive than PET, SPECT, and optical imaging (BLI and fluorescence), MRI has higher spatial resolution and is useful for real-time cell tracking and monitoring survival. MRI provides valuable simultaneous anatomical and physiological data. It is built on the principle that magnetic dipoles (unpaired nuclear spins) align in a magnetic field induced by a pulse from a radiofrequency coil. The radiofrequency coil then measures the return of the magnetic dipoles to their baseline spin orientation as an electromagnetic flux (1–100 MHz); the rate of return to baseline is the MR signal, and varies with the cellular environment . Modulation of the timing parameters of pulsing and recording is achieved primarily through T1 and T2 weighting, which affect image contrast. The ideal MRI contrast agent should be able to increase signal intensity without decreasing target cell activity, detect a small number of cells, and be retained by cells long enough and stably enough to permit longitudinal imaging . As with radionuclide imaging, MRI can employ both physical and genetic (such as -galactosidase and the transferrin receptor ) labels. The two most widely used physical labels are gadolinium chelates and superparamagnetic iron oxide (SPIO) nanoparticles. SPIO nanoparticles range from 30 to 180 nm in size. MRI protocols require that cells be labeled prior to transplantation. Once active intracellularly, SPIO nanoparticles decrease the contrast intensity of T2-weighted and T2*-weighted images .
SPIO particles have been used to track transplanted stem cells following myocardial infarction  and stroke , as well as the fate of migrating immune cells in cardiac allograft rejection . SPIO particles can be visualized for significantly longer time periods than radionuclide tracers. For example, SPIO-labeled MSCs were visualized in the swine myocardium following myocardial infarction for three weeks . Concerns remain regarding the safety and cellular toxicity of iron particles, though in vitro studies thus far have shown that MSCs labeled with ferumoxides-protamine sulfate complexes (Feridex) expressed normal lineage-defining surface markers and had no inhibited differentiation capacity . Another drawback is the difficulty in attributing signal attenuation to SPIO-labeled contrast agents vs. endogenous processes, as perivascular effects and the presence of hemosiderin can cause similar loss of signal on T2-and T2*-weighted MRI . The iron oxide label is diluted with each round of cell division, and, as with radionuclide approaches, decoupling of SPIO nanoparticles from the transplanted cells following cell death can lead to label uptake from surrounding cells or invading cells such as macrophages . As an example of this, while MRI detected signal from SPIO-labeled human peripheral blood-derived EPCs seven days post-injection into the adductor muscles of mice following ischemic injury by ligation of the femoral artery, cell death was confirmed by histology. The authors concluded that, despite early cell death, the EPCs were still able to exert an arteriogenic effect on ischemic tissue, but that MRI was an unsuitable methodology to track cell fate because the label was potentially taken up by invading macrophages .
Gadolinium chelates, in contrast to SPIO nanoparticles, increase signal intensity on T1-weighted images by inducing T1 relaxivity. As the first FDA-approved MRI contrast agent, gadopentate dimeglumine (Gd-DTPA) is used clinically in angiography . However, SPIO nanoparticles have some advantages over gadolinium chelates. For example, gadolinium chelates have been shown to decrease the proliferation of stem cells, whereas transplantation of SPIO-labeled vs. unlabeled cells has resulted in identical patterns of proliferation . The use of genetic labels to express MR reporter genes could overcome or mitigate these obstacles, but the potential adverse effect on genetic modification and the lower detection sensitivity of these MRI reporter genes (detailed in Table 2) remain problematic . While MRI does suffer from low detection sensitivity, physical labeling using SPIO nanoparticles and a balanced steady-state free precession (SSFP) sequence (3D-FIESTA) imaging system has permitted the visualization of single cells, whereas the majority of studies using radionuclide techniques detect populations in the range of 106 to 107 cells . Further research is needed to validate claims by studies that report such high MRI sensitivity. On a more practical level, MRI cannot be used in patients with pacemakers or other implantable devices such as defibrillators, due to the magnetic field’s interference with the device .
Bioluminescence imaging (BLI) requires incorporation of an optical reporter gene, most commonly the Firefly luciferase gene (Fluc, from the firefly Photinus pyralis). Light is produced when the enzyme oxidizes its substrate, D-luciferin, utilizing ATP as a cofactor, as well as oxygen and magnesium . Because the chemiluminescent reaction takes place only in living cells, bioluminescent signal is not emitted from non-viable cells. Additionally, light emission catalyzed by Firefly luciferase does not require an external light source, and the rapid enzyme turnover rate (t1/2 = 3 hrs) in the presence of substrate precludes enzyme accumulation and allows repeated measurements.
BLI has been used extensively in small animal studies to track and compare the survival, engraftment, and migration of a range of cell populations, including BMMNCs, skeletal myoblasts, and MSCs in the infarcted myocardium . In a hindlimb ischemia model, human embryonic stem cell-derived endothelial cells (hESC-ECs) preferentially localized to the ischemic hindlimb and incorporated into the microvasculature, and their localization (as assessed by BLI) was associated with improvement in limb perfusion (as assessed by laser Doppler) . This study underscores the relevance of study designs that incorporate molecular imaging techniques such as BLI to track the fate and localization of transplanted cells, together with non-cell-based functional outcome measures such as laser Doppler. In a direct comparison of SPIO nanoparticle labeling vs. Fluc reporter gene labeling of undifferentiated hESCs and differentiated hESC-ECs in the hindlimbs of immunodeficient mice, MRI showed similar signals from both cell populations for four weeks following transplantation. By contrast, BLI was able to distinguish between acute donor cell death for differentiated hESC-ECs vs. teratoma formation for undifferentiated hESCs  Fig. (2). Furthermore, BLI has been used to evaluate various immunosuppressive regimens for inducing long-term tolerance to xenogenic and allogeneic ESCs and induced pluripotent stem cell (iPSC) transplantation [81, 82]. Finally, insertion of the Fluc gene downstream of a tissue-specific promoter can lead to reporter activity upon cellular differentiation .
The chief disadvantage of BLI is that there is no human equivalent at this time. In small animal studies, it offers the ability to track cell fate longitudinally, with both high specificity and low background, but light transmission is a product of tissue depth and composition. Moreover, as a function of depth, signal attenuation and scatter are non-linear, adding to the difficulty of quantifying data when the model animal size increases . Finally, as with all genetic manipulations, epigenetic silencing can lead to loss of Fluc expression, which can be rescued in part by treatment with DNA methyltransferase inhibitor .
While fluorescent imaging (FLI) has been used by molecular biologists for decades, the past few years have seen novel applications of these technologies in vivo for the study of small animals. Fluorescent images are obtained by recording the shifted wavelength emitted from a subject receiving a specified wavelength of excitatory visible light. Green fluorescent protein (GFP), isolated from Aequorea victoria, emits green light (509 nM) upon excitation with violet light . Point mutations in the Aequorea-derived protein lead to excitation/emission-shifted variants. To highlight the range, far-red mPlum emits light in response to 590 nM excitation, while T-sapphire emits light in response to 399 nM excitation . Though traditional fluorescent imaging cannot obtain images at shallow tissue depths due to scattering of signal, the development of techniques such as fluorescence mediated molecular tomography (FMT) has overcome this barrier. FMT increases the number of source-detector pairs, resulting in low background noise and high, submillimeter resolution at a depth of 7.5 mm .
While traditional fluorescent imaging approaches use a genetic label, a new subset of fluorescent probes that use a physical label is emerging. Semiconductor quantum dots (QDs) detect molecules of interest in the cellular membrane by using fluorescent semiconductor nanocrystals . The excitation wavelengths can be manipulated to range from the ultraviolet through the near infrared range, and the emission spectra can be modulated by changing the probe size and composition. QDs are promising candidates for tracking cells in vivo due to their resistance to photobleaching and long-term fluorescent expression, allowing QDs to be distinguished from the autofluorescence of other cells when imaging with a longer exposure time . More recently, single-walled carbon nanotubes have been used in vivo as fluorophores to image deep tissues in the absence of excessive excitation and at a high frame rate . Prior to clinical adoption of QDs or nanotubes, future studies will need to address issues such as cellular toxicity from cytosolic aggregation, target specificity, and delivery techniques .
The primary disadvantage of fluorescent imaging is the high background signal that results from autofluorescence. Absorption of visible light by water, lipids, hemoglobin, and deoxyhemoglobin can also lead to false signal attenuation . As with SPIO nanoparticle labeling in MRI, the use of physical labels in QD imaging opens the possibility of label loss from the original cell and transfer to a nearby cell or to a macrophage following death of the donor cell. Such label decoupling and reuptake could lead to inaccurate quantification and lack of correlation between fluorescent signal and cell viability.
In addition to tracking the survival and integration of transplanted cells, a significant cohort of studies has imaged angiogenic and arteriogenic processes directly. To directly image non-cellular angiogeneic processes, a novel method has been validated that identifies tissues under hypoxic stress in atherosclerotic vascular disease. Intravenously-delivered 111In-labled VEGF121, which is secreted in response to hypoxia and binds to the VEGF receptors that are expressed at higher levels in ischemic tissue, was found to accumulate in greater levels in ischemic tissues compared to non-ischemic tissues in the rabbit hindlimb . Similarly, PET imaging showed that 64Cu-VEGF121 uptake in ischemic hindlimbs occurred through binding to VEGFR2, a VEGF receptor that plays a major role in regulating angiogenesis  Fig. (3).
αvβ3-integrin is expressed abundantly on the surface of proliferating endothelial cells and is involved in cell survival signaling and in cell migration, and has been used extensively as a target for the molecular imaging of the endothelium of newly-formed vessels . The first attempt to image angiogenesis, in a rabbit carcinoma model, used paramagnetic liposomes containing Gd3+ conjugated to an αvβ3-integrin antibody, and demonstrated the ability to directly target paramagnetic agents to the site of angiogenesis . Subsequent studies further established the utility of αvβ3-integrin targeting, and of multimodality imaging approaches. For example, a recent study investigated the efficiency of targeting liposomal nanoparticles functionalized with both an αvβ3-integrin-specific tripeptide sequence (for MRI detection) and with peptides specific for galectin 1 (for fluorescence detection), and demonstrated that dual-targeting resulted in a higher efficiency uptake of nanoparticles than occurred with a single target .
Several studies have used αvβ3-integrin targeting in contrast-enhanced ultrasound (CEU). CEU acoustically detects the production of gas-containing contrast agents or microbubbles, and has been applied experimentally to angiogenesis by altering the microbubbles’ chemical shell to target them to cells present at the site of angiogenesis (such as leukocytes), or by conjugating certain ligands to the microbubbles . In theory, single microbubbles can be detected. The aforementioned microbubble modifications have been used to target epithelial cell epitopes present during angiogenic processes; these modifications create a contrast that is detected by ultrasound imaging . Despite initial demonstrations of the utility of CEU, its clinical integration requires further refinement and understanding of the chemical and biological interactions of microbubbles.
CEU can also be used to measure perfusion changes and muscle flow reserve. Because patients with PAD have a significantly extended contrast medium wash-in time both at rest and during exercise compared to patients without PAD, CEU can be used to measure this parameter and to assess microcirculation before and after the initiation of cell therapy . Similarly, blood-oxygen-level-dependent (BOLD) imaging using MRI contrast has been applied to non-invasively measure oxygen delivery to hypoxic tissue. This technique is founded upon the increase in the ratio of oxyhemoglobin (which is diamagnetic) to deoxyhemoglobin (which is paramagnetic) as hypoxic tissue is perfused; patients with PAD have a decreased T2* signal and delayed maximum signal peak following postocclusive reactive hyperemia compared to healthy patients . Perfusion MRI techniques offer promise in the future to sensitively detect changes in oxygen delivery and microvasculature, but await further validation before clinical translation becomes feasible.
Micro Computed Tomography (MicroCT) is traditionally used to detect anatomical abnormalities, but has also been applied to the imaging of vasculature. Because the signal attenuation of small vessels does not differ significantly from that of surrounding tissue, the use of MicroCT to image vasculature requires the additional use of contrast agents and permits assessment of vascular or tumor volume, vessel density and branching patterns, and vascular surface area. As an example, MicroCT used in mice has shown that treatment with rPAI-123, an inhibitor of angiogenesis, can reduce tumor volume and neovasculature compared to treatment with saline .
Laser Doppler imaging has been used extensively to assess vessel patency and blood flow. For example, Laser Doppler imaging has been used to demonstrate a significant increase in blood flow following injection of BMMNCs into the ischemic limb, compared to controls receiving PBS . Nonetheless, the use of unlabeled cells severely limits in vivo analyses of cell fate. Traditional post-mortem fluorescence microscopy has shown that BMMNCs accumulate perivascularly in areas of collateral artery and capillary growth . However, this study and others that assess muscle perfusion would benefit from in vivo imaging modalities that offer temporal analyses of cell incorporation (or lack thereof) into vascular beds.
PAD, and its most severe manifestation as CLI, can significantly impair patients’ quality of life and is correlated with increased mortality, even in the absence of symptoms. At present, clinical interventions for degenerative arteriosclerotic disease aim to revascularize hypoxia-damaged tissues using either endovascular techniques or surgical bypass. Novel therapeutic approaches, however, are needed because traditional therapies have only limited success and exclude many patients. Preclinical data of stem cell therapy for the treatment of PAD indicate a promising alternative approach, particularly for patients ineligible for surgical intervention. Stem cell therapies have the potential to improve both functional (ankle-brachial index and transcutaneous oxygen pressure) and clinical (ulcer healing, pain-free walking distance, rest pain, amputation, and death rates) parameters. Results of the Phase III BONe Marrow Outcomes Trial in Critical Limb Ischemia (BONMOT-CLI) trial have not yet been published, but will provide insight into the efficacy of intramuscular bone marrow injection in patients whose only other option is major limb amputation . The phase II Intraarterial Progenitor Cell Transplantation of Bone Marrow Mononuclear Cells for Induction of Neovascularization in Patients With Peripheral Arterial Occlusive Disease (PROVASA) trial results have shown improved ulcer healing following intraarterial autologous BMMNC injection compared to placebo .
While stem cell therapies appear to be clinically effective, safe, and feasible, the mechanism by which transplanted cells support angiogenesis in ischemic vascular beds remains unclear. In particular, it is unclear whether the transplanted cells incorporate into the walls of new vessels, or if the cells themselves are dispensable and instead augment vessel formation by serving as a “cytokine factory” at the site of transplantation . Small animal studies utilizing longitudinal in vivo imaging to track the fate of transplanted cells have begun to offer mechanistic insights by correlating functional data with cell fate, but have been applied more uniformly to ischemic cardiac than to ischemic hindlimb models. The incorporation of imaging modalities, in both small animal studies and clinical trials, along with functional outcome measures in cell-based therapeutic efforts toward revascularization of ischemic limbs, should improve our understanding of the mechanisms underlying therapeutic angiogenesis.
Radionuclide imaging, including PET and SPECT, has already proved valuable clinically. While MRI cannot distinguish between viable and unviable cells, it does offer high spatial resolution and is particularly valuable for cellular localization. The fluorescent and bioluminescent reporter gene strategies offer both highly sensitive and quantifiable readouts of long-term cell survival and viability, but currently have no human equivalent and a relatively low spatial resolution (3–5 mm for bioluminescence; 2–3 mm for fluorescence), which is further limited by lack of tissue penetrance. Currently available technologies, as well as others in development, establish a collaborative synergy between clinical and basic science work. Developments over the past decades in cell-based therapy for the treatment of ischemic peripheral and cardiac disease have laid the foundation for improved clinical integration of cell labeling and imaging strategies. In vivo tracking of stem cells has improved the understanding of cell survival, migration, and localization following transplantation. At present, no single modality has the perfect combination of low toxicity, high sensitivity, and detailed resolution. The advantages and drawbacks of the four highlighted strategies must be taken into consideration when designing studies to track survival of transplanted cells, as well as to take advantage of their ability to promote regeneration of and incorporation into damaged tissues.
In the decades to come, the success of cell-based therapies to promote and evaluate vascular regeneration will rely even more on in vivo cellular imaging modalities. Strategies currently used to image the fate of transplanted cells will continue to improve in their spatial resolution, sensitivity, and longitudinal applications, while minimizing their negative effect on endogenous and induced regenerative processes. Clinicians and imaging laboratories must work together to develop techniques that can ask and answer relevant questions regarding how tracking and imaging cell-based therapies may best be used to improve our understanding of vascular regeneration, and to incorporate cell-based therapies clinically as an adjunct to traditional therapeutic interventions. Indeed, the pace of progress in the field will likely continue at an accelerated pace in the years ahead, bringing into clinics exciting new strategies for the visualization and evaluation of the regeneration made possible by cell-based therapies.
This work was supported in part by BWF CAMS, NIH HL099117, and NIH EB009689 (JCW).
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST