|Home | About | Journals | Submit | Contact Us | Français|
Recent technological developments in biomedicine have facilitated the generation of data on the anatomical, physiological and molecular level for individual patients and thus introduces opportunity for therapy to be personalized in an unprecedented fashion. Generation of patient-specific stem cells exemplifies the efforts toward this new approach. Cell-based therapy is a highly promising treatment paradigm; however, due to the lack of consistent and unbiased data about the fate of stem cells in vivo, interpretation of therapeutic remains challenging hampering the progress in this field. The advent of nanotechnology with a wide palette of inorganic and organic nanostructures has expanded the arsenal of methods for tracking transplanted stem cells. The diversity of nanomaterials has revolutionized personalized nanomedicine and enables individualized tailoring of stem cell labeling materials for the specific needs of each patient. The successful implementation of stem cell tracking will likely be a significant driving force that will contribute to the further development of nanotheranostics. The purpose of this review is to emphasize the role of cell tracking using currently available nanoparticles.
The practice of medicine represents a unique relationship between the health professional and the patient [1–3]. Thus, a personalized approach has always been the essence of health care delivery. The patient’s demographics, family history, the results of blood and radiological exams and of biopsies have been employed for years to tailor a specific therapy for each patient. That is why the personalized aspect of medicine is continuously being emphasized, and has recently gained more and more interest.
The term “Personalized Medicine” has recently acquired a new dimension as a result of the explosion of data about biomedical processes. Anatomical and physiological data on a molecular level can now be generated for each individual patient. It has long been anticipated that biotechnology will revolutionize the medicine of the future . We live at a time when this is becoming a reality. Indeed, the term “Personalized Medicine” was coined because of the advances in clinical pharmacogenetics [5–7] and was popularized by Hood, a visionary doctor and researcher, who invented the DNA and protein sequencer and synthesizer . Hood developed the concept of “4×P” medicine, which is short for “Predictive, Preventive, Personalized, and Participatory medicine” . However, such a novel concept, based on using the entire breadth of data available about each patient to make therapy more individualized, clashed with the current classical approach of evidence-based medicine (EBM) . EBM focuses on using the best evidence from randomized clinical trials (RCTs) about treatment efficacy. This process requires dichotomization of patients into experimental and control groups, resulting in large heterogeneity within groups . However, RCTs remain the mainstream approach and, as of today, from a legal point of view, they are mandatory to obtain Food and Drug Administration approval for the marketing of therapeutic products . One criticism of RCTs is that a selected pool of “ideal” patients is recruited, which ignores complex clinical situations. As a result, the outcomes of RCTs are relevant to only a fraction of the patient population . Increasing discontent with EBM is leading to growing support for the concept of personalized medicine . As a compromise between the two trends, the new term “personalized EBM,” has recently been coined .
Whatever the approach, it would seem that personalized medicine will be a definite factor in the practice of medicine, and we must choose to use it wisely to save lives and improve the quality of life for individual patients. The current developments in biotechnology provide us with an unprecedented opportunity to practice personalized medicine in the most effective way. Thus, it is our obligation to develop approaches that meet the current demands of clinical practice and to use personalized medicine to improve the treatment of all patients .
Although personalized medicine arose from pharmacogenetics , it has now infiltrated many fields of healthcare [18–20]. Despite the advances in pharmacology, it became apparent that many disorders were beyond the capabilities of current treatment methods . As a response to widely unmet clinical needs, an approach based on using cells for therapy has emerged [22,23]. The discovery and characterization of stem cells, with self-renewal and differentiation capacities, also accelerated this field, making regenerative medicine a new independent discipline . However, embryonic stem cells are plagued by ethical and immunorejection problems, and a shortage of adult stem cells precludes their wide clinical application. Only the recent discovery of induced pluripotent stem cell technology, which enables any cell of the body (including skin fibroblasts) to be changed into a pluripotent stem cell, with the potential to differentiate toward any cell type in unlimited quantities, may make the advent of large-scale cytotherapeutics possible . Of 106 ongoing and planned neuro-restorative clinical trials, nearly two-thirds focus on cell-based therapy . The possibility to derive, repair, propagate, and transplant cells specifically for each individual patient takes personalized medicine to an entirely new dimension .
A Swedish group, who pioneered cell-based therapy for Parkinson’s disease in their proof-of-concept clinical neurotransplantation trials, carefully selected and treated individual patients, thus highlighting the importance of a personalized approach [23,28,29]. However, traditional RCTs later performed in the USA failed to reveal a statistically significant positive effect of cell therapy for Parkinson’s disease. Although there were positive effects in some patients, other patients experienced side effects that exacerbated the disease [30,31]. Thus, the EBM approach, which neglects the heterogeneity of response to treatment within groups, may be too simplistic an approach for use with stem cell therapy. The personalized approach, however, seems more likely to succeed with the use of stem cell treatments.
Bone marrow reconstitution by stem cells is a well-established therapy for malignancies [32,33] and aplasias [34,35], and these are excellent examples of the success and impact that cell-based therapy has had on medicine. In contrast, the treatment of diseases that affect many other organs and tissues with stem cell therapy is still in its infancy and requires further development . New technology, such as non-invasive imaging of transplanted cells to monitor their fate in vivo, will be instrumental in developing effective clinical applications for cell therapy [37,38].
There is a wide array of diseases awaiting effective treatment, with neurological disorders being a prominent example of cases where treatment options are scarce, and stem cell-based therapies offer hope for scores of patients. The inconvenience of life-long injections for the treatment of diabetes with inevitable, long-term complications makes diabetes another application where cell replacement therapy is very attractive, especially in the light of the increasing incidence of this disease . Stem cell therapy for diabetes and liver failure with cadaver-derived cells resulted in some positive effects ; however, limited access to those sources precludes widespread clinical application . The effective production of hepatocytes from induced pluripotent stem cells (iPSC)  and from embryonic stem cells (ESC) [43,44] has been reported, and that may address the problem of donor material shortage. In addition to the hepatocyte replacement approach, the administration of bone marrow stem cells enhanced the functional hepatic reserve after extended hepatectomy . It has also been shown that transplanted bone marrow cells may counteract the complications of anti-cancer therapy by restoring long-term fertility  and helping to repair the intestinal mucosa . Connective tissue diseases, including disc and joint disorders, occur frequently in young adults, causing long-term absence from work, and forcing premature retirement, and, as a consequence, major socio-economic challenges [48,49]. The positive effects of cell therapy for musculoskeletal injuries were demonstrated initially in race-horses [50,51]. This was followed by the spectacular therapeutic success in Bartolo Colon, a New York Yankee baseball pitcher who suffered from a shoulder injury, but who could return to this sport after stem cell therapy (May 14 2011, The Wall Street Journal). There has also been extraordinary progress in the use of stem cells for the ex vivo production of connective tissue structures, such as the trachea, and subsequent successful transplantation in patients .
Stem cell-based therapy is very promising for many disorders, with some encouraging outcomes already reported. However, the results are still variable, making the likelihood quite low that insurance companies and national health systems will cover such procedures, at least in the nearest future.
The tremendous progress in material science and biotechnology has enabled unparalleled miniaturization and manipulation of materials at the molecular or even atomic scale. By definition, a nanotechnology deals with materials that range from 1 to 100 nm in size in at least one of three dimensions. Fabrication at the nanoscale level fundamentally multiplies the repertoire of attainable products and offers products of unprecedented properties. There are two main approaches to nanofabrication: top-down and bottom-up .
The top-down approach is based on using nano-machine tools to shape matter. This lies within the domain of physicists, and one of the most spectacular examples was the weaving of a pattern from single atoms using atomic force microscopy (AFM) . This type of nanofabrication has also found its way to surgery with the concept of nanosurgery using multi-photon microscopy .
The bottom-up approach is based on the self-assembly of atoms or molecules toward nanostructures . In other words, the principle of this approach is to make a non-covalent bond art out of covalently bond blocks. Although this concept seems new, it has already been employed in nature, with the most apparent manifestation that of complex protein structures self-organizing from a simple amino-acid sequence. This type of nanofabrication is now being widely investigated, primarily by chemists, to create functional, highly ordered, hierarchical structures. There is a wide diversity of self-organizing structures:
The techniques for fabrication, with atomic precision, continue to increase; thus, further advances toward high-throughput manufacturing are on the horizon. The above-listed nanostructures have found numerous applications in virtually all fields of science and medicine, and also proved to be well-suited for stem cell labeling and non-invasive, in vivo tracking of stem cells.
The diversity of nanomaterials that have recently become available opens the opportunity to tailor these materials to the specific needs of individual patients . While the proof-of-principle of personalized nanomedicine has been reported, applying these novel tools clinically is not trivial . Adding stem cells to this equation makes things even more complex . Introducing cell-based therapy as a new potent strategy in the therapeutic arsenal has been met with great enthusiasm, but putting this therapy into routine clinical practice proved difficult. Studies about the transplantation of stem cells of various origins have been conducted for years, initially in animal models and then in patients, offering hope for effective treatment. However, there are severe problems with the interpretation of the outcomes of stem cell-based treatments. One of the leading obstacles is the lack of consistent and unbiased data about stem cell distribution and survival in vivo. Typically, for drug development, the pharmacokinetics and pharmacodynamics are routinely employed to characterize the basic features of the tested drugs in the organism . The benefits of any surgical intervention are usually confirmed through post-operative imaging. These methods offer a conclusive assessment and allow a judgment about the efficacy of pharmacotherapy or surgery. Such standards, unfortunately, have not yet been developed for stem cell therapy. Although neuroimaging was strongly supported as a base for rational stem cell therapy , to date, only a few clinical studies have followed this recommendation . The resistance to the implementation of cellular imaging into clinical protocols may stem from the notion that transplanted stem cells will undergo self-guidance in terms of their migration and functional integration, which is unrealistic at best, or may be due to the lack of robust, versatile, and validated methods by which to monitor the fate of transplanted stem cells.
Nanotechnology now offers many solutions for labeling cells and those are well-suited for personalized medicine. With cellular imaging, cell transplantation-based therapy may be tailored to the individual patient rather than to a large heterogeneous population. Such an approach, which would consider the pathology of each specific patient, would help to maximize the therapeutic effect. While cell therapy for internal disorders, such as diabetes, liver failure, or myocardial infarction, can be evaluated with more objective measures, neurological disorders are the most challenging in this regard. Although a meta-analysis of preclinical results revealed the positive effects of cell therapy in neurological disorders , there is still no consensus about the mechanisms that mediate cell-dependent effects, or about timing, dosage, type of cells, and other factors . Thus, stem cell tracking is expected to play a pivotal role in defining these mechanisms, and this is essential for further optimization of cell therapy .
There are many imaging modalities used in clinical practice and the choice of a modality is determined by the specific diagnostic question, availability, and cost-effectiveness. Additional factors that determine the suitability of an imaging method are specificity, sensitivity, resolution, and radiation exposure of individual modalities. Imaging plays a critical and constantly growing role in medicine and there are well-developed imaging algorithms that are routinely used to evaluate the course of many diseases. Consequently, it is of the utmost importance that a new imaging methodology, such as cell tracking, does not interfere with some of the existing critical imaging algorithms. The goal is not only to see transplanted cells, but also to image pathological processes within the organ of interest. Some of the available cellular imaging techniques do interfere with anatomical or functional imaging, such as in the case of iron oxide in MRI. Thus, the selection of an imaging modality should be driven by the need for a particular anatomical or functional imaging technique required for evaluation of the underlying pathology. This concept is termed the “imaging window,” and could be different in each patient, given existing co-morbidities and other factors.
Magnetic resonance imaging is capable of producing images of the human body with an unsurpassed detail. It has also been shown, with iron oxide as a cell label, that MRI is an excellent method for detailed demonstration of the cell location after transplantation. In one reported study in cancer patients, iron oxide-labeled cells were injected into lymph nodes, and imaging showed a striking mis-injection rate, with off-target delivery in half the cases . Due to the superior homogeneity of the magnetic field within the brain, MRI has also been successfully used for neurotransplantation ; however, it was shown that iron oxide may interfere with MR imaging parameters . Since various MR imaging techniques, such as functional MRI (fMRI), diffusion tensor imaging (DTI), or perfusion MRI, may be necessary to evaluate the progression of central nervous system (CNS) diseases or the healing process, alternative tracers, including those suitable for other imaging modalities, are desirable. While there is an effort to develop new generation “switch-able” MRI contrast agents, their sensitivity is still below clinical utility . Nevertheless, for many cell therapy applications, especially in cases when a precise reference to anatomy is critical, MRI of iron oxide-labeled cells is still the method of choice.
X-ray CT is characterized by excellent temporal resolution, very high spatial resolution, and satisfactory anatomical and topographical depiction with relatively low soft tissue contrast. As such, it is a potentially interesting candidate modality for stem cell imaging, particularly for applications in the brain or lungs. Well-developed methods for image co-registration  enable spatial superposition of images acquired using CT and MRI, and allow complementary multimodality imaging. CT is the leading modality for the evaluation of bone damage and repair[81,82], thus it would be desirable to follow the transplanted cells for bone healing with a different modality so as not to interfere with the CT evaluation of osteogenesis. Compared to MRI, stem cell tracking with CT is far less developed. However, it has been shown that gold nanoparticles can be used to image a molecular target  and/or cells  in vivo. Labeling of mesenchymal stem cells with gold nanotracers was found to be safe, with no detrimental effects on cell function . In addition to standard X-ray CT scans, it has recently been reported that this modality can produce multi-color images [86,87] that potentially could be used to monitor several different processes simultaneously, including tracking two populations of cells.
Nuclear medicine is characterized by an excellent in vivo sensitivity and whole-body imaging capabilities. However, the “hot spot” imaging of radionuclides usually requires additional scanning with another anatomy-providing imaging modality, followed by image co-registration and superimposition. Luckily, integrated PET/CT scanners are now widely available. In addition, co-registration of MRI and CT is well-established and those images can readily be fused with PET results . Integrated PET/MRI scanners are another option. The integration of PET and high-field scanners is a major challenge, as this result in interference, with a decreased quality of images in both modalities . Although there has been significant progress in the development of hardware for the purposes of nuclear medicine, the applicability of radionuclide-based cell tracking is still unclear, primarily due to the issues of safety and toxicity. While systemic injection of radionuclides results in negligible whole-body radiation, the situation is quite different for cell labeling. A relatively high accumulation of radionuclides in single-labeled cells and the prolonged (days/weeks) irradiation that concentrates most of the energy within the same cell may be detrimental to stem cells . The superb signal-to-noise ratio may warrant labeling only a fraction of the transplanted cells, but even those can be at risk of mutation and subsequent tumorigenic transformation. Thus, radionuclide-based stem cell tracking in patients should be introduced cautiously. Another limitation of this modality is the short half-life of radioactive tracers, which do not extend one week, even with the longest-lived isotopes. While the clinical application of nuclear medicine for stem cell tracking seems problematic, the application of this modality in the preclinical setting may still be valuable.
The major advantages of USG are the low cost, wide availability, and lack of toxicity. USG may be easily employed to monitor and navigate the device used for cell injection , but visualization of the cells themselves is complex. As in other modalities, transplanted cells require labeling with a contrast agent. A typically used and commercially available contrast agent for USG consists of air bubbles. Functionalized bubbles, injected intravenously, have been used to detect subcutaneously transplanted cells . The bubbles have been found to be taken up by tumor cells, improving tumor visualization with USG . However, the major limitation of USG is associated with the severe attenuation of ultrasound waves by bones, which practically eliminates USG as a robust method for cell tracking within the CNS.
Due to light scattering and absorption by the tissues, most applications for optical imaging are in small animals, where bioluminescence imaging, in particular, is intensively exploited to monitor the status of transplanted cells [94–96]. The clinical application of optical imaging, for obvious reasons, is largely limited to superficial targets. One very important application for this approach is based on fluorescent probes that target tumor cells used to guide brain tumor resection in patients. Relatively good tissue penetration by photons in the near infrared wavelength region makes optical probes of these characteristics most promising for broader applications in patients. The so-called “optical window” that offers good tissue penetration is limited by water absorption of the photons above 1300 nm, and hemoglobin absorption below 700 nm . Currently, near-infrared (NIR) spectroscopy is widely used in studies of brain oxygenation  or even for fMRI-like assessment of resting-state functional connectivity in the brain . The low cost and portability of near infrared (NIR) scanners contributes to the rapidly growing interest in this modality. Recent advances in nanocrystal formation, with exceptionally high photoluminescence within a near infrared range, have been shown to be very promising in small animal experiments, and may also find applications in the clinical setting .
Stem cell tracking has been introduced as a response to the need to address distinct scientific questions, including the evaluation of cell distribution, cell survival, and cell function. Stem cell tracking is being applied in a variety of disorders and with many different transplantation routes .
The evaluation of cell distribution following transplantation seems to be an essential, basic task, and such as pharmacokinetics for drug development should be recommended for any cell therapy-based study. Determining the location of transplanted cells within the recipients’ body is critically important, and the value of this information increases the farther the distance between the site of the cell deposit and the desired cell destination. The evaluation of cell distribution is usually most important within the first few days after transplantation, and until cells reach the final destination [103,104]. However, in some circumstances, cells may migrate over longer periods of time [105–107]. The intravenous route for cell delivery, with a high degree of uncertainty about cell distribution, is especially appealing for cell tracking [108–111]. The intra-arterial route is more targeted in terms of cell delivery, it also can suffer from mis-injection, which, however, can be easily captured by in vivo cell tracking (Fig. 1) . This approach also provides information about the dynamics of cell redistribution or with-in the body. In animals, cell distribution can also be assessed by labor-intensive post-mortem histopathology, but this is not an option in the clinical setting. Monitoring the time course of cell delivery allows an evaluation of the desired and actual cell locations. With access to that information, the interpretation of experimental/clinical data is much easier. As shown in reports on cell-based therapy trials, the response to cell therapy is often variable, and that variability may, at least in part, result from less than optimal cell delivery in some patients . The response to therapy may also reveal that cell distribution is inadequate or insufficient, and this may account for a suboptimal outcome. In this context, all aspects of cell delivery and cell distribution must be carefully addressed to maximize the chance for optimal cell delivery to the correct destination.
An appropriate assessment of cell distribution requires the application of a high-resolution, tomographic modality, such as MRI, or, alternatively, a multimodality approach that can co-register high-sensitivity low-resolution and low-sensitivity high-resolution acquisitions (i.e., PET and MRI) .
It has been shown recently that cell tracers used for non-invasive imaging may persist in tissues following cell death , or these tracers can be diluted by cellular proliferation . Thus, monitoring cell viability is a more challenging task than was thought originally . In small animals, reporter gene-based (bioluminescence) imaging has been found to be extremely useful, but, as mentioned above, it is not clinically applicable , and, as of now, no clinically applicable solution has been found by which to monitor cell viability.
There have been continuous efforts to establish a gene reporter system for MRI and several initial prototype genes have been developed[118,119], but current approaches are still plagued by insufficient sensitivity and specificity . An alternative to reporter genes would be a “smart probe,” producing contrast based on cell viability status. Most desirable probe would be one that produces positive signal within living cells and no signal upon a cell death.
Direct imaging of the function of cells in vivo can, at present, only be done using quite invasive approaches, such as two-photon microscopy. Non-invasive imaging techniques still lack the sensitivity and resolution to directly image cell function. However, indirect imaging of transplanted cell function has long been used [122,123] to obtain information about cell function. A good example of such indirect monitoring is related to a very specific situation in Parkinson’s disease, where well-defined dopaminergic transmission is compromised by the disease process, and probes targeted to the dopaminergic machinery can visualize its function. While such a functional imaging paradigm has been used to evaluate the efficacy of cell-based therapy, it is not capable of providing a definitive answer about whether recorded dopaminergic transmission stems directly from the transplanted cells, or indirectly, from endogenous repair mechanisms. Nevertheless, indirect measures of grafted cell functionality are an important guide to enable accurate judgments about graft functionality and facilitate further optimization of the therapeutic protocol.
The field of stem cell research benefits greatly from nanotechnology, which offers many solutions for stem cell tracking. Nanotechnology provides a variety of direct stem cell labeling options, as well as facilitates the use of other important methods, such as reporter genes. Stem cell tracking techniques can be divided, based on the placement of contrast material, into intracellular, with contrast agent localized within the cell, and extracellular, where signal is derived from the out-side of a cell, but is closely related to the cell. Stem cell tracers may be detectable by just one modality (unimodal) or by several modalities (multimodal).
Internalization of iron oxide particles by specific cell populations was observed as early as 80 years ago and was used for the magnetic selection of Kupffer cells . After the introduction of MRI to the clinic, it was reported that facial cosmetics used for mascaras and tattooing resulted in artifacts on MR images, and iron oxide particles were found to be responsible for this effect[125–127]. After coating by polymers to increase biocompatibility and protect aggregation , iron oxide particles were used for MR imaging of the liver to diagnose malignancy . There is now a variety of sizes of iron oxide particles available for cell labeling, ranging from micron-sized  to nanometer-sized particles . Iron oxide nanoparticles are one of the most frequently used agents for cell tracking.
The first application of iron oxide nanoparticles to label cells other than Kupffer macrophages was the ex vivo labeling of peripheral blood mononuclear cells . That was followed by studies on the labeling and imaging of transplanted neural cells [133–136]. Of note is the fact that the size of iron oxide nanoparticles determines the efficiency of cell up-take, with much higher labeling efficiency for small nanoparticles ca. 70 nm diameter (SPIO), than for ultra-small iron oxide nanoparticles (USPIO) . One of the SPIO formulations (Feridex®) approved by the Food and Drug Administration (FDA) in 1996 as a liver contrast agent, following the development of methods that facilitated the labeling of non-phagocytes, including stem cells, was extensively used and characterized in preclinical research, with the intention to use Feridex® clinically thereafter [138–140]. Unfortunately, in 2008, the company stopped manufacturing Feridex®, citing economic among other reasons, and leaving the field without a clinically approved SPIO agent. Recently Feraheme®, a new formulation of iron oxide, has been approved by the FDA for the treatment of anemia. Due to their ultra-small size (USPIO), labeling cells with these nanoparticles is more challenging; however, recent reports demonstrated a labeling technique for Feraheme® that makes them detectable by MRI [141,142]. A follow-up study demonstrated improved labeling with the use of self-assembling nanocomplexes by combining Feraheme®, heparin, and protamine, yielding larger, ca. 100 nm-size complexes (Fig. 2), facilitating stem cell up-take . Of note is that all these agents are FDA-approved for other clinical uses.
Due to the very strong MRI signal, iron oxide enables almost microscopic visualization of cells, especially when imaged using a high-field MR scanner. However, long-term observation has revealed that SPIO within the tissue and contrast on MRI persists despite the death of transplanted stem cells [96,144,145]. In this context, SPIOs offer robust and specific information about the distribution of labeled cells; however, long-term monitoring of the fate of stem cells requires cautious interpretation. For applications in peripheral organs, dual-modality MRI/PET is an attractive option with SPIO-based MRI reporting on the detailed location of cells and PET imaging of thymidine kinase products reporting on cell viability . Such a strategy for imaging within the CNS is, however, complicated by the lack of penetration of certain PET probes across the blood-brain barrier.
Amid the reported limitations of iron oxide nanoparticles, there have been continued efforts to improve the properties of these agents with a multitude of new surface modifications/coatings, revealing better labeling and toxicity profiles. Poly(N,N-dimethylacrylamide)-coated maghemite , aminosilane coating , amine surface modification , carboxymethyl chitosan modification , 1-hydroxyethylidene-1.1-bisphosphonic acid (HEDP) coating , non-toxic protein transduction domain (PTD) conjugation , higher density carboxyl groups modification , and D-mannose-modification  were used for this purpose. The versatile and tunable coating strategies for iron oxide nanoparticles , may enhance the customization of labeling for individual applications and the requirements of individual patients.
It has been shown that the contrast mechanism of iron oxide nanoparticles in MRI depends on the particle size. While SPIOs were traditionally used to provide negative T2 contrast, the ultra-small iron oxide nanoparticles, in addition to the T2 contrast, can also produce the more-desirable positive T1 contrast . The large-scale synthesis of uniform and extremely small-sized iron oxide nanoparticles (ESION) for positive T1 contrast may increase the utility of iron oxide as a contrast agent .
Another use for iron oxide nanoparticles, beyond providing out-standing MR contrast, is based on their magnetic characteristics that enable the guidance and navigation of labeled cells after transplantation. This can be accomplished with the use of external magnets placed over the target organ, which results in increased stem cell homing to desired organs [158–160].
Gadolinium (Gd) is a T1 contrast agent that generates positive contrast, thus avoiding difficulties in the interpretation of signal voids, which often complicate imaging with SPIOs . However, clinically used gadolinium chelates do not produce a positive T1 signal after internalization by the cells . The widespread use of gadolinium as a cell labeling agent may be hampered by difficulties in the synthesis of stable, water soluble solutions . However, recently, there have been several reports on new approaches that incorporate gadolinium within the metal-based core of nanoparticles , or polymer-based scaffolds . While Gd chelates are characterized by a relatively low signal per molecule, the formation of gadolinium oxide nanoparticles with certain magnetic properties has been shown to be useful for tracking hematopoietic cells . With the advent of commercially available dextran-coated gadolinium oxide nanoparticles—e.g. Gado CELLTrack (BioPAL)—it is expected that this approach may be further pursued . Further work to optimize the surface coatings of gadolinium oxide nanoparticles is under way, with a range of modifications, such as using shells made of PEG  and DEG, silica , and albumins . Another approach is to capture gadolinium atoms within carbon cages made of nanotubes or fullerenes . However, the cardinal issue that would determine the advantage of gadolinium-based nanoparticles over SPIOs is their clearance after the death of grafted cells. Recently, it has been reported that a free Dex-DOTA-Gd3+ contrast agent clears from the transplantation site, as opposed to iron oxide particles; however, the credibility of this result would have been significantly improved by the presentation of raw MR images and not merely volume-rendered scans, as presented in this publication . Most importantly, prolonged retention of gadolinium will results in release of free Gd3+ ions, which are toxic. Hence, it is very unlikely that Gd-based cell tracking will reach the clinic.
The similarity of manganese and gadolinium, with regard to providing T1 contrast in MRI, has been reported some time ago . However, the in vivo [176,177] and in vitro  neurotoxicity of manganese favored gadolinium as a widely clinically applicable contrast agent . Notably, it has been recently reported that the toxicity of manganese to cells other than neurons is low, comparable to that of iron, and much lower than that of silver and molybdenum oxide . An efficient method for the production of manganese oxide nanoparticles, followed by coating with a biocompatible PEG-phospholipid shell, sparked interest in manganese as an MRI contrast agent . This manganese oxide nanoparticle formulation was successfully used to image transplanted cells, but was characterized by a lower sensitivity that was inferior to that of iron oxide . A modification toward mesoporous silica-coated hollow manganese oxide nanoparticles yielded increased T1 signal through better access of water molecules to the magnetic core .
Fluorine is a chemical element that is practically absent in the human body; thus, 19F imaging can be used to produce contrast devoid of any tissue background. Another advantage of this method is that, during the same imaging session, both 1H anatomical and 19F “hot spot” images can be acquired and superimposed , providing data in a quantitative manner. While the signal-to-noise ratio (SNR) of 19F is about 89% that of 1H per nucleus, the fluorine-based contrast agent requires a very high density of 19F nuclei, similar to that of 1H density in tissues. This has been accomplished by synthesizing fluorine compounds from hydrocarbons through the exchange of 1H to 19F nuclei. Even prior to the era of MRI, fluorocarbons were used as an X-ray contrast agent , or as a blood substitute . Thus, the biocompatibility of fluorocarbons is undoubted. Perfluoropolyether nanoparticles have been successfully used to track dendritic cells , pancreatic islets , and neural stem cells (Fig. 3) [192,193]. While perfluorocarbon is a very promising contrast agent for “hot spot” MRI, its utility is still limited by the low sensitivity of the method , requiring a significant amount of cells for in vivo imaging .
Chemical exchange saturation transfer (CEST) is a new contrast mechanism based on the exchange of protons between certain molecules and bulk water as the result of saturation with an off-resonance pulse sequence, or the use of rare earth metals that provide a paramagnetic shift in the resonance frequency (PARACEST). The advantage of PARACEST contrast is that it is saturation pulse frequency-specific, and thus, completely switch-able, and contrast is generated only after applying a specific saturation pulse. As such, CEST does not interfere with other MR parameters, i.e., T2, T1, or diffusion. The sensitivity of CEST for proteins and polymers is very low; however, the use of PARACEST has been shown to produce a stronger signal. Chelates of europium and ytterbium were shown to be good PARACEST agents, and these chelates, incorporated into dendrimer nanoparticles, have been used in vivo . PARACEST nanoparticles, as yet, have not been used specifically for stem cell tracking, but this technology is ready to be tested for cell transplantation pending sufficient sensitivity of detection.
Heavy metals have long been used as contrast agents for electron microscopy [197–199]. The biocompatibility of gold has already been established in dentistry[200,201] and the low cytotoxicity of cell labeling with gold nanoparticles has been confirmed in vitro . Just a few years ago, it was reported that gold nanoparticles were characterized by very high X-ray absorption, far beyond iodinated compounds, confirming that these particles made good X-ray contrast agents, when systemically delivered in small  and large animals . Only recently, successful tracking of cells labeled with gold nanoparticles using micro-CT has been shown . As discussed above, gold has many features advantageous for cellular imaging; however, a significant disadvantage is its cost. Thus, there was an effort to replace it with low-cost, bio-inert tantalum oxide nanoparticles, which reportedly have the many advantages of gold nanoparticles, including low toxicity [205,206].
The term “quantum dots” has been coined for three-dimensionally confined semiconductor nanocrystals . These nanocrystals are characterized by a stable and unique fluorescence emission band that is easily tunable by manipulating the size or the composition of crystals . The broad absorption and narrowemission bands facilitate multicolor imaging of dynamic cellular events. The production of quantum dots with an emission spectrum in the near infrared range contributes to their use for in vivo imaging[210,211], including successful tracking of stem cells transplanted into the brain [212,213]. It has been shown that near infrared quantum dots (QD 800) produce a higher in vivo signal compared to far red quantum dots (QD 655) . Toxicity studies have demonstrated that quantum dots have no effect on the viability, proliferation, or differentiation potential of stem cells ; however, pro-thrombotic adverse events were reported following intravenous transplantation of labeled mesenchymal stem cells .
Highly fluorescent, π-conjugated, polymer-based semiconductor nanoparticles are an organic alternative to quantum dots . These nanoparticles are characterized by a very high emission rate, minimal “blinking” behavior, and excellent photostability. Various polymers emit photons of distinct wave-lengths, which enable multicolor imaging. Recent advances toward the construction of near-infrared π-conjugated polymers herald alternative strategy to quantum dots . The successful synthesis inorganic-organic/nanocrystals-polymer hybrids with unique properties has also been reported, which may further enhance the detection of transplanted cells .
Among the diversity of biolabels, fluorescent nanodiamonds are also worth mentioning. They have not yet been used for in vivo tracking of stem cells, but their characteristics, such as photostability, chemical non-reactivity, biocompatibility, and emission in the far-red bandwidth, make them another interesting nanoparticle for non-invasive imaging . The atomic compositionof nanodiamonds enables enhancement by fluorescence resonance energy transfer (FRET), but at the cost of the introduction of another label into the cells [221,222]. Nanodiamonds were shown to be relatively safe for the cells [223,224]; however, a slight increase in DNA repair proteins in nanodiamond-labeled embryonic stem cells has been reported .
While near infrared fluorescent nanostructures proved to be useful for in vivo imaging, the autofluorescence is a factor that limits the sensitivity of this method. The process of UCL may help with that limitation. The phenomenon of UCL takes place when absorption of light by the molecule results in the emission of photons with a higher energy, but a shorter wavelength, described as the Anti-Stokes emission process. This phenomenon has been known for many years, but only recently have nanoparticles with such properties, suitable for in vivo applications, been produced. Since there are few, if any, endogenous biological materials that display an upconverted emission, the images obtained by the use of such materials are completely devoid of back-ground . Rare-earth nanophosphors, with upconversion within the near infrared range (both absorption/emission), were used as photoluminescent probes for small animal in vivo bioimaging . Tissue phantoms provided proof-of-principle for cellular imaging, using NaYF4 nanocrystals doped with Tm3+, with detection reaching 3 mm in depth . Very recently, the replacement of yttrium with lutetium, resulting in cubic sub-20 nm NaLuF4-based upconversion nanophosphors, demonstrated a dramatic increase in signal, enabling in vivo imaging of large animals, such as rabbits and pigs .
For years, echogenic bubbles have been used as an ultrasonic contrast agent [230,231]. The second harmonic wave provided a background-free signal using an ultrasonic device. Since the frequency of the second harmonic wave was dependent upon the bubble diameter, the size of the bubbles had to be tuned to the frequency of clinical ultrasonic devices. Based on these calculations, the dimension of bubbles suitable for this application would be in the range of micrometers, and products of this size are being manufactured and are commercially available. However, these preparations are actually oversized and not suitable for intracellular labeling.
Although standard ultrasonography equipment cannot be used to detect cells labeled with nanobubbles, it has been shown that nanobubble-labeled liver cells are detectable by ultrasound microscopy . However, the exposure of the cells to nanobubbles was related to toxicity from some of the bubble formulations; thus, their implementation into clinical practice should be undertaken with caution . Nevertheless, the progress in the formation of stable nanobubbles in a well-controlled manner [235,236] may facilitate the introduction of ultrasonography as a viable strategy for cellular imaging.
Several substances are characterized by their production of contrast that is detectable by more than one imaging modality. Such approach with exploiting one nano-structure for multimodal imaging is particularly useful, with only few reported applications.
Europium-doped mesoporous gadolinium oxide nanoparticles, in addition to MRI signal, are also characterized by strong phosphorescence, which is useful for optical imaging, and these nanoparticles have already been applied to the labeling of HeLa cells . Doping of gadolinium oxide with La3+ ions provides an additional upconversion luminescence effect .
While iron oxide is one of the best MR contrast agents in terms of signal change, there are metals with even higher magnetic moments, potentially further increasing the signal. The high magnetic moment of cobalt nanoparticles was shown to be even more efficient in signal generation than iron oxide, but, due to insufficient stability, these particles were not useful for biomedical applications . However, manufacturing a hollow cobalt and platinum (CoPt) alloy overcame the hurdle of stability without the accompanying signal decrease. Such magnetic CoPT nanoparticles were successfully employed for stem cell labeling . The presence of platinum as a component of this CoPt alloy renders these particles X-ray visible. It has also been shown that CoPt nanoparticles can be prepared using a dendrimer matrix .
Radionuclides are frequently used for stem cell imaging. Due to the very high sensitivity of radionuclides, only a minimal amount of label is necessary, and chemical complexes such as In111-inoxine can be used for cell tracking over the time until the decay of radioactivity, which, in the case of In111 is several days . Nanotechnology contributed much to the field of stem cell research with the introduction of multimodal nanoparticles, and by the addition of radionuclides to existing nanoparticles. Since the relatively fast decay process precludes the storage of the radionuclides, the synthesis of such nanoparticles should be as straightforward as possible to enable on-site synthesis immediately prior to their application. To meet these requirements, rapid, size-controlled synthesis of dextran-coated 64Cu-doped iron oxide nanoparticles has been developed using simple microwaving .
Perfluorooctylbromide, in addition to the MRI signal of fluorine, is also characterized by strong X-ray absorption due to the presence of bromide atoms, making it suitable for CT imaging . Perfluoropropane nanobubbles, in turn, are used for ultrasound imaging and are also visible on 19F MRI .
While there are only a few examples of tracers made from homogenous multimodal nanoparticles, there are numerous hybrid nanoparticles employed for multimodal stem cell tracking (Table 1). Usually, these hybrids are composed of two or more substances used for unimodal nanostructures, or with one nanostructure and an additional optical dye or/and radionuclide. The core of such a unimodal nanoparticle is typically surrounded by a shell made of another unimodal material. An iron oxide core coated with a gold shell exemplifies the types of nanoparticles designed for dual MR and CT imaging [246,247]. A more advanced design includes a CT-visible core made of perfluorooctylbromide with a conjugated europium chelate that produces a CEST MRI signal . Fluorescein or rhodamine  bound to gadolinium nanoparticles enable both magnetic resonance and optical imaging. An attractive in vivo stem cell tracking system for clinical translation would be the integration of magnetic nanoparticles with near infrared quantum dots. Attaching Cy5.5 to iron oxide nanoparticles also enables both MR and NIR imaging [252,253]. Dendrimer-based nanoparticles can be coated with various labels, including gadolinium and Cy5.5 . Triple-modality nanoparticles, with a heavy metal core coated with paramagnetic and optical tags, have also been described (Fig. 4).
Reporter genes have been found to be very useful for identifying transplanted cells . The application of reporter gene-based imaging methods allows for unambiguous conclusions about the viability of transplanted cells. In contrast to many nanoparticles, reporter proteins produced by engineered cells are very short-lived and degrade rapidly following cell death, with a resulting loss of signal. They are also resistant to contrast dilution during cell proliferation . The most popular reporter genes that are suitable for in vivo imaging in small animals are based on light-emitting proteins or enzymes. The recent discovery of reporter genes with fluorescent protein emitting in the near-infrared range may open up applications for optical imaging in large animals or even in humans . However, the clinical application of reporter genes for stem cell tracking faces some challenges, including the safety issues associated with manipulating the genome or the effects of gene introduction into the cells. While the in vivo gene therapy of malignant tumors [257,258] or the ex vivo introduction of reporter genes to therapeutic immune cells  have been used, the genetic alteration of stem cells in regenerative medicine, due to their expected long-term survival, has met with much more resistance. Efficient gene delivery to human therapeutic stem cells is still challenging. While transfection/transduction of immune cells or immortalized human stem cells is quite efficient, the genetic modification of primary human cells with therapeutic potential is very difficult. While reporter gene-expressing therapeutic cells can be derived from transgenic 849 donor animals, such a strategy is not applicable for human cells. Thus, there is an effort to improve reporter gene delivery to primary stem cells using developments in nanotechnology. Nano-coating is a very interesting strategy to enhance gene delivery to the cells. The surface coating of individual cells with an amino-acid functionalized calcium phosphate nano-film was initially used with the intention of guiding the process of cell differentiation. The so called “exo-scaffolds,” characterized by a positive charge of hydroxyapatite, were found to significantly enhance non-viral transfection methods, probably by increasing plasmid concentration at the cell surface . Nano-coating with a chitosan and chitosan-hyaluronan was shown to increase the endocytosis of both DNA plasmids and iron oxide nanoparticles .
The technology of microencapsulation has been developed primarily to protect transplanted cells against an immune attack by the host . Since visualization of transplanted cells is necessary to correctly determine the cell-dependent effect, it has been proposed that the capsules containing these cells might also be monitored with in vivo imaging. For many years, alginate has been used to encapsulate living cells . The co-encapsulation of pancreatic beta islet cells in alginate capsules with iron oxide [264–266] or barium (sulfate) [267–269] enabled tracking by MRI and CT, respectively. Then, trimodal gadolinium-gold capsules were introduced, with detectability by MRI, CT, and ultrasonography (Fig. 5) . This was achieved by the functionalization of gold nanoparticles with gadolinium chelates and through cross-linking of alginate with clinical-grade protamine sulfate . As some of the nanoparticles with beneficial imaging parameters may be toxic, a strategy has been developed to shield cells from such particles using capsule-in-capsule (CIC) methods, in which nanoparticles are first encapsulated within an internal capsule, and then the cells are incorporated within the external capsule layer, providing a physical separation of cells and nanoparticles . By adding perfluorocarbons to the capsules, a trimodal capsule composition can also be synthesized that can be detected with 19F MRI, CT, and USG .
Another approach is to label the capsule wall, as for iron oxide and gadolinium chelates . Particles incorporated into the wall of the capsule demonstrated a sufficient signal for in vivo MRI, and did not expose the cells to the label itself. PEG-lipids used for single-cell encapsulation of HEK293 cells (Fig. 6)  and hepatocytes  were conjugated with FITC for in vitro visualization. A chitosan-alexa fluor 647 conjugate has been employed for the in vitro imaging of multilayer chitosan/alginate nano-coating of beta islets . A modification of this strategy with the use of near infrared dye could potentially be applicable to clinical capsule imaging.
Nanogels are composed of cross-linked polymer nanoparticles and have been increasingly used as injectable scaffolds for transplanted cells, especially useful in case of cell deposition within cavities. The use of scaffold materials facilitates uniform cell distribution and provides a substrate for cell adhesion [279,280], as well as promoting cell differentiation and survival . It has been shown that the deposition of cells within a hydrogel enhanced the therapeutic effects of the cells in a model of acute kidney injury  and myocardial infarction[283,284]. Imaging of nanogels has also been explored, with reports on labeling with fluorescence tags, such as rhodamine B  or quantum dots [286,287] or magnetic iron oxide . This strategy with tagging gels that encapsulate therapeutic cells could potentially provide information about whether deposited cells remain within the targeted site or redistribute. Nanogels were also shown to improve the efficiency of cell labeling with quantum dots [289,290] or gold nanoparticles .
The growing interest in the application of nanotechnologies to cell tracking requires tools by which to evaluate the interactions of nanotracers with living cells at the subcellular level . Coherent anti-Stokes Raman scattering (CARS) microscopy enables the visualization of metal oxide particles based only on their electron density, obviating the need for combination with other labels for in vitro visualization [293,294]. Further advances with the application of stimulated Raman scattering provides such visualization at high temporal resolution, enabling recording at video-rate . Surface-enhanced Raman spectroscopy enabled the imaging of the cellular transport pathways of gold nanoparticles (Fig. 7) . This allows confirmation, in living cells, that the nanoparticles were efficiently internalized, and in which cell compartment these nanoparticles are located.
The interest in stem cell therapy is keen, and, over the last decade, has increased dramatically; a search of the clinicaltrials.gov database on 03/24/2012, using the key word “stem cells,” returned a stunning 3,971 records. As presented in this review, there is a multitude of techniques for tracking cells, some suitable only for the laboratory setting, but some with clinical translation potential. To date, only 19 clinical trials (which included a total of 145 patients) have reported the use of stem cell tracking . Nuclear medicine was a dominating modality with the majority of applications in the field of cardiology. Contrast materials used for these studies were either 111-indium and 99-technetium chelates or 18-FDG. Thus, all nuclear medicine applications did not use nanotechnology approaches.
Iron oxide nanoparticles have been the only nanostructures used in a clinical setting [298–301]. Considering the abundance of nanomaterials, the low rate of clinical translation is somewhat surprising and disappointing, particularly if patients could greatly benefit from more targeted and personalized stem cell therapies. Further research on the biocompatibility and safety of the most promising agents should help in the rapid translation of new nanotechnology developments.
While the enthusiasm related to the biomedical applications of nanotechnology is substantial, these approaches are not devoid of risk, which requires careful planning and preclinical testing before wide clinical application. Some nanomaterials that were envisioned to be highly useful therapeutically proved toxic, including such agents as carbon nanotubes , as well as silver  and copper nanoparticles [304,305]. The toxicity may even depend on the size of the material, as was shown for gold nanoparticles with intraperitoneal injection. In this instance, 17 nm particles led to impaired cognition in mice, while such an effect was not observed after the application of larger 37 nm gold nanoparticles . Gold nanoparticles are known to generate heat after exposure to near infrared light  or ionizing radiation , which is actually used in cancer therapy, but which may interfere with diagnostic imaging. Transplanted stem cells are meant to function over extended periods of time; thus, the evaluation of potential toxic effects should be performed for both immediate and long-term effects. It has been shown that even if no acute side effects are observed, some particles may cause an inflammatory response during long-term observation , inhibit chondrogenesis [310,311], or temporarily interfere with the speed of migration .
It is not only the safety, but also the reliability of stem cell tracking methods that should be considered. The interpretation of in vivo imaging can be complicated by such processes as contrast dilution following cell division [115,313], uptake of the label by phagocytes following the death of transplanted cells, or, as has been recently reported, transfer of magnetic nanoparticles via microvesicles due to stress (starvation) .
Current treatment methods for many diseases are ineffective and inadequate, thus driving the growing interest in stem cell therapy. The implementation of personalized medicine, based on extensive and accurate individual patient data, opens up the possibility to design therapy more specifically, customizing it to the needs of each particular patient. This approach heralds a more tailored and likely more efficient use of stem cells. The ability to non-invasively, monitor the fate of transplanted cells, in real-time, is a crucial requirement to conduct experimental therapies, and only that ability can provide meaningful data that will allow the advancement of cell therapy. The advent of nanotechnology boosts the arsenal of methods for tracking stem cells in patients. Thus, investigators and clinicians who introduce cell therapies into the clinical realm should be knowledgeable and able to briskly navigate between the available methods of cell tracking so as to customize the protocol for each individual case. In general, the selection of the imaging modality and stem cell tracer or label should be based on the primary disorder, but co-morbidities must also be considered, which may influence which modality is best for the individual patient so as to avoid interference with the imaging techniques necessary to evaluate disease progression or recovery. The condition of the patient is also an important factor, as a worse prognosis may justify a treatment strategy that is riskier but has the potential to provide greater benefit, such as cell therapy with the use of cell tracking. At this early stage of the field of cell therapy, randomized trials, because of the inherent within-population heterogeneity, do not seem to be the best strategy. Rather, a personalized approach to treatment appears more appropriate.
The power of personalized nanomedicine lies in the diversity of nanomaterials, which enable individualized tailoring of materials to the specific needs of each patient. The successful implementation of stem cell tracking will likely be a significant driving force that will contribute to the further development of nanotheranostics .
The authors are supported by 2RO1 NS045062 JWMB, RO1 DA026299, U54 CA151838, MSCRFII-0193, MSCRFII0052, and a Kolumb Fellowship from the Foundation for Polish Science (MJ). The authors thank Mary McAllister for editorial assistance.
This review is part of the Advanced Drug Delivery Reviews theme issue on “Personalized nanomedicine”.