Human embryonic stem (hES) cells are pluripotent stem cells capable of self-renewal and differentiation into virtually all cell types. Thus, they hold tremendous potential as cell sources for regenerative therapies. The concurrent development of accurate, sensitive, and noninvasive technologies capable of monitoring hES cells engraftment in vivo can greatly expedite basic research prior to future clinical translation. In this study, hES cells were stably transduced with a lentiviral vector carrying a novel double-fusion reporter gene that consists of firefly luciferase and enhanced green fluorescence protein. Reporter gene expression had no adverse effects on cell viability, proliferation, or differentiation to endothelial cells (human embryonic stem cell-derived endothelial cells [hESC-ECs]). To compare the two popular imaging modalities, hES cells and hESC-ECs were then colabeled with superparamagnetic iron oxide particles before transplantation into murine hind limbs. Longitudinal magnetic resonance (MR) imaging showed persistent MR signals in both cell populations that lasted up to 4 weeks. By contrast, bioluminescence imaging indicated divergent signal patterns for hES cells and hESC-ECs. In particular, hESC-ECs showed significant bioluminescence signals at day 2, which decreased progressively over the following 4 weeks, whereas bioluminescence signals from undifferentiated hES cells increased dramatically during the same period. Post-mortem histology and immunohistochemistry confirmed teratoma formation after injection of undifferentiated hES cells but not hESC-ECs. From these data taken together, we concluded that reporter gene is a better marker for monitoring cell viability, whereas iron particle labeling is a better marker for high-resolution detection of cell location by MR. Furthermore, transplantation of predifferentiated rather than undifferentiated hES cells would be more suited for avoiding teratoma formation.
Disclosure of potential conflicts of interest is found at the end of this article.
Human embryonic stem (hES) cells are derived from the inner cell mass of preimplanted blastocysts . They have been shown to differentiate into a variety of cell types that represent endoderm, ectoderm, and mesoderm origins via three-dimensional structures called embryoid bodies (EBs), which at least partially mimic the spatial organization of the embryo [1, , –4]. Various lineages have been derived from human and mouse ES cells, including neurons , cardiomyocytes [6, 7], hematopoietic cells [8, 9], osteogenic cells , hepatocytes , insulin-producing cells , keratinocytes , and endothelial cells [3, 4, 14]. Furthermore, these cells appear to be weakly immunogenic, with absent major histocompatibity complex (MHC)-II and only low levels of MHC-I molecules . Given their unlimited self-renewal and pluripotency capacity, hES cells represent a new and exciting avenue for stem cell therapy. In cell culture, hES cells can differentiate into endothelial cells through successive maturation steps [4, 9, 16]. Therefore, the isolation and use of hES cell-derived endothelial cells (hESC-ECs) have potential therapeutic applications, including cell transplantation for repair of ischemic tissues and tissue-engineered vascular grafts.
Stem cell therapy is an exciting area of research that promises future treatment of many diseases . However, to fully understand the beneficial effects of stem cell therapy, investigators must be able to track the biology and physiology of transplanted cells in living subjects over time. At present, most cell therapy protocols require histological analysis to determine viable engraftment of the transplanted cells. The development of sensitive, noninvasive technologies to monitor this fundamental engraftment parameter will greatly aid clinical implementation of cell therapy. In reporter gene imaging, stem cells can be genetically engineered to express various reporter genes before transplantation. The reporter genes can be detected by sensitive imaging devices such as the optical coupled device, single-photon emission computed tomography, or positron emission tomography (PET) . Magnetic resonance (MR) imaging of stem cells is also an emerging application for monitoring cell engraftment. In particular, stem cells labeled with superparamagnetic iron oxide (SPIO) particles can be identified in vivo as hypointensities in MR images as the iron shortens transverse proton relaxation times. The spatial and temporal resolution of MR imaging allows the location of iron-labeled donor cells to be monitored noninvasively over several weeks in vivo [19, –21]. A fundamental drawback of hypointense signal is the difficulty in distinguishing iron-labeled cells from the surrounding air, hemorrhage, necrosis, and macrophages. To address these problems, techniques to generate positive contrast to enhance signal- and contrast-to-noise ratios of the iron-labeled stem cells have been developed. Cunningham et al. have reported off-resonance (OR) MR imaging of iron-labeled mouse embryonic stem (ES) cells to generate positive contrasts through suppression of background tissue . However, whether these different methodologies can be applied for studying the engraftment of hES cells and their derivatives in vivo has yet to be examined. To help answer these important questions, we performed a head-to-head comparison of tracking undifferentiated hES cells and differentiated hESC-ECs using reporter gene and iron particle imaging.
Materials and Methods
Maintenance and Differentiation of Human Embryonic Stem Cells
Undifferentiated hES cells (H9 line [WiCell Research Institute, Madison, WI, http://www.wicell.org]; passages 35–45) were grown on an inactivated mouse embryonic fibroblast (MEF) feeder layer as previously described . Briefly, the cell was maintained at an undifferentiated stage on irradiated low-passage MEF feeder layers on 0.1% gelatin-coated plates. The medium was changed daily and consisted of Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 medium (F12), 20% knockout serum replacement, 0.1 mM nonessential amino acids, 2 mM l-glutamine, 0.1 mM β-mercaptoethanol, and 4 ng/ml rhFGF-2 (R&D Systems Inc., Minneapolis, http://www.rndsystems.com). The undifferentiated hES cells were treated with 1 mg/ml collagenase type IV in DMEM/F12 and scraped mechanically on the day of passage. Prior to iron labeling and endothelial differentiation, hES cell were seeded onto Matrigel (BD Biosciences, San Diego, http://www.bdbiosciences.com)-coated plates in CM prepared from MEF as follows . MEF cells were harvested, irradiated with 50 Gy, and cultured with hES medium without basic fibroblast growth factor (bFGF). CM was collected daily and supplemented with an additional 4 ng/ml bFGF before feeding hES cells.
In Vitro Differentiation of hESC-ECs
To induce hES cell differentiation, undifferentiated hES cells were cultured in differentiation medium containing Iscove's modified Dulbecco's medium and 15% defined fetal bovine serum (FBS) (HyClone, Logan, UT, http://www.hyclone.com), 0.1 mM nonessential amino acids, 2 mM l-glutamine, 450 μM monothioglycerol (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 50 U/ml penicillin, and 50 μg/ml streptomycin, in ultralow-attachment plates for the formation of suspended EBs as previously described [4, 9]. Briefly, hES cells cultured on Matrigel-coated plate with CM were treated with 2 mg/ml dispase (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) for 15 minutes at 37 °C to loosen the colonies. The colonies were then scraped off and transferred into ultralow-attachment plates (Corning Enterprises, Corning, NY, http://www.corning.com) for EB formation.
Whole-Mount Immunostaining of Human Embryoid Bodies
Whole-mount immunohistostaining of human embryoid bodies (hEBs) was performed as previously described  with minor modifications. The hEBs were fixed in methanol and dimethyl sulfoxide (4:1) at 4 °C overnight. For staining, the rehydrated hEBs were first blocked by two incubations in 2% bovine serum albumin and 0.2% Tween-20 in phosphate-buffered saline (PBSBT) and then with PBSBT containing mouse anti-human monoclonal antibody CD31 (BD Biosciences) overnight at 4°C. The hEBs were washed five times in PBSBT each for 1 hour at 4 °C for the initial three washes and at room temperature for final two. The primary antibody was labeled by incubating the hEBs with Alexa 594-conjugated goat anti-mouse IgG (Invitrogen) in PBSBT overnight at 4 °C and nuclear-counterstained with 4,6-diamidino-2-phenylindole (DAPI).
Flow Cytometry Sorting of hESC-ECs
Single-cell suspensions from day 12 of differentiated hEBs were obtained by treatment with 0.56 units/ml Liberase Blendzyme IV (Roche Diagnostics, Indianapolis, http://www.roche-applied-science.com) at 37 °C for 20–30 minutes. Cells were passed through a 40-μm cell strainer . Cells were incubated for 30 minutes at 4 °C with mouse phycoerythrin (PE)-conjugated anti-human CD31 (BD Biosciences). The CD31+ cells were isolated using a FACScan (Becton Dickinson). To generate hESC-ECs, the isolated CD31+ cells were grown on 0.1% gelatin-coated or 4 μg/cm2 human fibronectin (Calbiochem, San Diego, http://www.emdbiosciences.com)-coated plates in EGM-2 with 5% FBS. The medium was changed every 2–3 days.
Biological Characteristics of hESC-ECs
Flow cytometry analysis, DiI-ac-low-density lipoprotein (LDL) uptake assay, and Matrigel assay were used to confirm endothelial cell phenotype within these CD31+ purified hES cells. Antibodies used in this study were PE-conjugated anti-CD31 (BD Pharmingen) and allophycocyanin (APC)-conjugated anti-KDR (R&D Systems), APC-conjugated anti-mouse IgG2a, and mouse anti-human VE-cadherin (BD Pharmingen). The stained cells were analyzed using a FACSVantage (Becton Dickinson). Dead cells stained by propidium iodide were excluded from the analysis. Isotype-identical antibodies served as controls (BD Pharmingen). Approximately 5 × 105 undifferentiated hES cells and 2 × 105 differentiated hESC-ECs were used to run the fluorescence-activated cell sorting (FACS). For DiI-ac-LDL uptake assay, hESC-ECs were incubated with 10 μg/ml DiI-Ac-LDL (Molecular Probes, Eugene, OR, http://probes.invitrogen.com) at 37 °C for 6 hours. After being washed twice with phosphate-buffered saline (PBS), the slides were fixed and counterstained with DAPI, mounted with mounting medium (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com), and covered with coverslips prior to detection with fluorescence microscopy as described . The formation of endothelial tubes was assessed by seeding cells in 24-well plates coated with Matrigel and incubating them at 37 °C for 12 hours as described [14, 24].
Quantitative reverse transcription (RT)-polymerase chain reaction (PCR) assays were performed using the human endothelial cell biology RT2 Profiler PCR Array (SuperArray Bioscience Corporation, Frederick, MD, http://www.superarray.com) on an ABI Prism 7900 HT (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Data analysis is available at the company website (http://www.superarray.com/pcr/arrayanalysis.php). Briefly, total RNAs were isolated using RNeasy (Qiagen, Waltham, MA, http://www1.qiagen.com) from undifferentiated hES cells at day 0, differentiated hEBs at day 12, hESC-ECs (after CD31 sort), and human umbilical endothelial cells (HUVECs) as positive control. First-strand cDNAs were generated using iScript Select cDNA Synthesis Kit (Bio-Rad, Hercules, CA, http://www.bio-rad.com). For real-time PCR, first-strand cDNAs were added to the RT qPCR Master Mix (SuperArray Bioscience). Samples were heated for 10 minutes at 95 °C and then subjected to 40 cycles of denaturation at 95 °C for 15 seconds and annealing and elongation at 60 °C for 1 minute. Methods for conventional RT-PCR analysis and primer sequences for endothelial-specific genes (CD31, VE-cadherin, KDR, Oct-4, GAPDH) are described in the supplemental online Methods and the supplemental online Table, respectively.
Lentiviral Transduction of hES Cells with Double-Fusion Reporter Gene
To track transplanted cells in vivo, hES cells were transduced at multiplicity of infection of 10 with self-inactivating lentiviral vector carrying a human ubiquitin promoter driving firefly luciferase and enhanced green fluorescence protein (Fluc-eGFP). Stable clones were isolated using FACS for eGFP expression. Afterward, Fluc activity within different cell numbers was confirmed ex vivo using a Xenogen IVIS 200 system (Xenogen, Alameda, CA, http://www.caliperls.com) as described . Nontransduced hES cells (control) and hES cells with the double-fusion (DF) reporter gene (hESC-DF) were stained for Oct-4 (Chemicon, Temecula, CA, http://www.chemicon.com). The undifferentiated hES cell colonies were fixed in 4% paraformaldehyde in PBS for 15 minutes. Nonspecific binding was blocked with 4% normal goat serum for 30 minutes, following which the colonies were stained with antibodies to Oct-4, incubated with Alexa 594-conjugated rabbit anti-goat secondary antibodies (Invitrogen) for 30 minutes, and nuclear-counterstained with DAPI. Images were obtained with a Zeiss Axiovert microscope (Carl Zeiss Micro Imaging Inc., Thornwood, NY, http://www.zeiss.com). Subsequently, the processes for in vitro endothelial cell differentiation and characterization were the same as those for control nontransduced hES cells described earlier.
Labeling of hES Cells and hESC-ECs with Iron Particles
Transduced hES cells and derived hESC-ECs were labeled with Feridex IV-protamine sulfate (FE-Pro) as described  with minor modifications. The commercially available SPIO suspension (Feridex IV; Berlex Laboratories, Inc, Wayne, NJ, http://berlex.bayerhealthcare.com/index.html) contains particles approximately 80–150 nm in size and has a total iron content of 11.2 mg/ml. Preservative-free protamine sulfate (American Pharmaceuticals Partner Inc., Schaumburg, IL, http://www.appdrugs.com/), 10 mg/ml, was prepared as a fresh stock solution of 1 mg/ml in distilled water at the time of use. SPIO at a concentration of 100 μg/ml was put into a mixing tube containing serum-free culture medium. Protamine sulfate (12 μg/ml) was added, and the entire suspension was mixed for 5–10 minutes. The final FE-Pro suspension was added directly to the existing medium, incubated overnight. The final concentrations of Feridex IV and protamine sulfate were 50 and 6 μg/ml of medium, respectively. After overnight incubation, the hES cells and hESC-ECs were washed twice with PBS and harvested by treatment with collagenase IV and trypsin, respectively. Next, the trypan blue exclusion assay was used to assess viability and cytotoxicity of iron labeling on hES cells and hESC-ECs. Six samples were performed and averaged for these assays.
Transplantation of hES Cells and hESC-ECs into Murine Hind Limbs
All procedures were performed on 8–10-week-old female SCID beige mice (Charles River Laboratories, Wilmington, MA, http://www.criver.com) (n = 15) according to the Stanford University Animal Care and Use Committee guidelines. Mice were anesthetized with inhaled isoflurane (2%–3%). Approximately 1 × 106 undifferentiated hES cells and 1 × 106 hESC-ECs (both stably transduced with DF reporter gene and colabeled with SPIO particles) were injected into right and left hind limbs of the same mouse, respectively, in 100 μl of PBS.
Optical Bioluminescence Imaging of Transplanted Cell Fate in Living Mice
Bioluminescence imaging was performed using the Xenogen IVIS 200 system. After intraperitoneal injections of reporter probe D-Luciferin (150 mg of luciferin/kg), animals were imaged for 2 seconds to 2 minutes. The same mice were scanned for 4 weeks. Imaging signals were quantified in units of maximum photons per second per square centimeter per steridian as described . Bioluminescence imaging was performed by an investigator blinded to the study conditions (X.X.).
MR Imaging of Transplanted Cell Fate in Living Mice
Afterward, in vivo MR imaging was performed on the 1.5 T-MR scanner (Signa; GE Medical Systems, Milwaukee, http://www.gehealthcare.com/euen/mri/index.html) using a 3-inch surface coil. The mice were anesthetized with intraperitoneal injection of ketamine (100 mg/kg) and xylazine (20 mg/kg) and imaged in a prone position. After three-dimensional-plane localization, images were acquired in a coronal plane using a gradient-recalled echo (GRE) sequence (repetition time [TR] = 100 milliseconds; echo time [TE] = 10 milliseconds; flip angle = 30°; field of view [FOV] = 8 × 8 cm; slice thickness = 1 mm; slice gap = 0 mm; matrix = 256 × 256; number of excitations [NEX] = 4). Multiple contiguous coronal slices consisting of 9–10 slices were acquired for complete coverage of the mice hind limb. Following GRE, OR sequence was applied to the transplanted labeled cells. Briefly, OR uses spectrally selective radiofrequency pulses to excite and refocus the off-resonance water surrounding the labeled cells while suppressing on-resonance signal. This generates positive contrast from the hydrogen protons adjacent to the labeled cells and helps identify and estimate the volume of the labeled cells . Projectional OR imaging was performed with TR = 800 milliseconds, TE = 14 milliseconds, FOV = 20 cm, and matrix = 256 × 128 with two discrete Fourier transformation encoding and an 8-millisecond readout . Measurement of signals in the region of interest was performed with the ImageJ software. The signal volume of GRE images was measured by summation of hypointense area of each slice. The OR signal enhancement area was obtained by measuring the area of projectional OR image. MR imaging was performed by an investigator blinded to the study conditions (Y.S.). The protocol for ex vivo MR imaging of cell suspension is included in the supplemental online Methods.
Immunohistochemical Staining for Macrophages, Iron Particles, and Various Cell Types
Mice were euthanized at 4 weeks after cell transplantation. Both hind limbs were embedded in paraffin and cut into 5-μm sections. H&E staining and Prussian blue staining were carried out to identify the fate of transplanted hES cells, and hESC-ECs were doubly labeled with iron particles and reporter genes. To examine whether the iron particles are localized in the macrophages, staining of macrophages (Mac-3) and iron particles was performed. Sections were incubated first with the monoclonal rat anti-mouse macrophage Mac-3 antibody (BD Pharmingen) for 1 hour and then with a second antibody, biotinylated goat anti-rat secondary antibody (Invitrogen), for another hour. Streptavidin-horseradish peroxidase (HRP) was then applied for 30 minutes, and HRP activity was detected with 3,3′-diaminobenzidine. For iron particle staining, sections were incubated with Perls reagent in the dark for 30 minutes and counterstained with fast red. For immunofluorescence histology, both hind limbs were embedded in O.C.T. compound (Sakura Finetek Inc., Torrance, CA, http://www.sakuraeu.com), and 5-μm frozen sections were cut. For immunostaining, sections were incubated with primary antibodies, rat anti-mouse macrophage Mac-3 antibody (BD Pharmingen), rabbit anti-green fluorescent protein (anti-GFP) (Invitrogen), and goat anti-mouse CD31 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), followed by incubation with secondary antibody, APC-conjugated mouse anti-rat IgG (BD Pharmingen), Alex 488-conjugated donkey anti-rabbit IgG (Invitrogen), and Alex 594-conjugated donkey anti-goat IgG (Invitrogen), and then counterstained with DAPI. Histologic interpretation was performed by a pathologist blinded to the study conditions (A.J.C.).
Analysis of variance (ANOVA) and repeated measures ANOVA with post hoc testing, as well as the two-tailed Student's t test, were used. Differences were considered significant at p < .05. Unless otherwise specified, data are expressed as mean ± SD. H&E staining and Prussian blue staining were carried out to identify the fate of transplanted hES cells and hESC-ECs doubly labeled with iron particles and reporter genes.
Characterization and Differentiation of hES Cells
To compare the fate of undifferentiated versus differentiated hES cells in vivo, we first established a differentiation protocol to induce hESC-ECs (Fig. 1A). Undifferentiated hES cell (H9 line) were maintained either on irradiated MEF feeder cells or on Matrigel-coated plates in the presence of MEF condition medium . To isolate endothelial cells from hES cells, undifferentiated hES cells cultured on Matrigel-coated plated were placed into Petri dishes with differentiation medium for induction of EB formation. Previous studies have shown that differentiated hEBs from hES cells contain endothelial cells that can be isolated by CD31  or CD34 markers . In this study, whole-mount immunostaining confirmed that within day-12 hEBs, the CD31+ endothelial cells were organized in specific channel-like structures (Fig. 1B). These data confirm that hES cells cultivated as hEBs spontaneously differentiated to endothelial cells and formed blood vessel-like structures. Initial FACS analysis showed that ∼3% of cells expressed CD31 marker. This population of CD31+ was sorted, cultured, and propagated for further analysis to confirm endothelial cell traits. In particular, we examined CD31, VE-cadherin, and kinase insert domain receptor (KDR), which are known to be markers for endothelial differentiation of ES cells [3, 4, 9, 26, 27]. As expected, both FACS and quantitative RT-PCR analysis indicated that undifferentiated hES cells expressed endothelial markers at a low level. After endothelial induction, intercellular adhesion molecule (ICAM) 1, CD31, vascular cell adhesion molecule (VCAM) 1, von Willebrand factor (VWF), and VE-cadherin were upregulated, whereas Oct-4 expression (marker for undifferentiated state) was downregulated (Fig. 1C, 1D; supplemental online Fig. 1A). Quantitative RT-PCR analysis showed similar endothelial gene expression pattern of hES-ECs compared with HUVECs but great disparity between hESC-ECs and hES cells (Fig. 1D; supplemental online Fig. 1B, 1C). Interestingly, KDR was continuously expressed in both undifferentiated and differentiated hES cells as shown by RT-PCR and FACS (Fig. 1C, 1D; supplemental online Fig. 1A). This pattern of expression is unlike that of mouse ES cells whereby KDR is specifically expressed on hemangioblasts but not on undifferentiated ES cells . After 1 week of culturing, flow cytometry was again performed using antibodies directed against endothelial markers such as VE-cadherin and CD31. Isolated cells continued to express robust levels of both VE-cadherin and CD31 (>80%). These CD31 cells morphologically resembled HUVECs and can form tube-like structures on Matrigel (Fig. 1E).
Genetic Labeling of hES Cells with DF Reporter Gene
To develop an imaging platform for tracking transplanted hES cells and hESC-ECs in living animals, we used a DF reporter gene consisting of Fluc-eGFP (Fig. 2A). The efficiency of self-inactivating lentiviral vector for transducing hES cells was ∼20% (data not shown). Both control nontransduced hES cells and stably transduced hES cells showed similar expression patterns of stem cell markers Oct-4 on immunostaining, suggesting minimal side effects by reporter gene on maintaining stem cell state (Fig. 2B). Upon culturing onto 24-well plates, we also observed a strong correlation (r2 = 0.99) between Fluc activity and cell numbers ex vivo using the Xenogen IVIS system (Fig. 2C). The cell proliferation and cell viability data are also similar between control hES cells and transduced hES cells (data not shown). Overall, these data are consistent with our previous studies showing minimal effects of reporter genes on mouse ES cell survival, proliferation, and differentiation [18, 28, 29]. In addition, the differentiation procedure and differentiation efficiency of stably transduced hES cells were similar to those of control nontransduced hES cells. Isolated cells express robust levels of both VE-cadherin and CD31 and avidly incorporate DiI-ac-LDL (Fig. 2D, 2E). Uptake of DiI-ac-LDL has been used to characterize mature endothelial cells . The isolated cells similarly can form tube-like structure on Matrigel (data not shown). Overall, reporter gene expression did not seem to significantly affect the typical characteristics of endothelial cells as measured by immunostainings, DiI-ac-LDL uptake, and Matrigel angiogenesis assay.
Physical Labeling of hES Cells with Iron Particle
To track transplanted cells using MR imaging in vivo, hES cells and hESC-ECs (both stably expressing the DF reporter gene) were colabeled with iron particles. After staining with Prussian blue, iron-labeled cells were found to have cytosolic accumulation of iron. Interestingly, the iron uptake of hES cells was less than hESC-ECs, which is likely due to the difference in their physical size (Fig. 3A). Similarly, in vitro cellular MR imaging demonstrated that dephasing signals in hES cells were weaker than those in hESC-ECs. Representative in vitro cellular magnetic resonance imaging (MR imaging) of labeled cells is shown in Figure 3B. Finally, the cell viability assay showed no significant difference between control unlabeled hES cells and iron particle-labeled hES cells (Fig. 3C).
Longitudinal MR Imaging of hES Cell Survival in Living Animals
To evaluate whether MR imaging can be used to serially monitor cell survival following transplantation, we injected 1 × 106 iron-labeled hES cells into left hind limbs and 1 × 106 iron-labeled hESC-ECs into right hind limbs of SCID beige mice. We then imaged these iron-labeled cells repetitively for up to 4 weeks. Representative serial MR images of iron-labeled cells by GRE and OR  sequences during the 4-week period are shown in Figure 4A and 4C, respectively. Control animals injected with unlabeled cells showed no MR signals, as expected. Similar to the in vitro data showing more cytosolic accumulation of iron particles within the larger hESC-ECs (Fig. 3A), the data from the GRE and OR signals were also noted to be stronger in the right (hESC-ECs) compared with left (hESC) hind limbs. Four weeks after cell transplantation, MR imaging with GRE sequence showed a significant increase in the physical size of the left hind limb injected with iron-labeled hES cells, which is due to teratoma formation (Fig. 4A). However, when normalized to day 2, analysis of both MR imaging signals (GRE and OR) produced no difference at 4 weeks, as shown in Figure 4B and 4D (p = .001).
Longitudinal Reporter Gene Imaging of hES Cell Survival in Living Animals
Since the same animals described above were injected with hES cells and hESC-ECs double-labeled with iron particles and Fluc-eGFP, we decided to analyze the ability of reporter gene imaging to assess stem cell fate (e.g., engraftment, proliferation, and death). Longitudinal bioluminescence imaging was also performed in the same animals for 4 weeks (Fig. 5A). In both hind limbs, bioluminescence signals were most robust immediately after transplantation (day 2). In the hESC-EC group, bioluminescence signals progressively decreased from day 2 to day 28 (p = .001 vs. control for all time points), indicating acute donor cell loss. Control animals injected with PBS showed no imaging signals, as expected. Quantitative analysis shows that the hESC-ECs survival activity at day 21 was less than 1.5% compared with day 2. This survival pattern of differentiated hESC-ECs is markedly different compared with undifferentiated hES cells, which was >40-fold higher compared with day 2 (Fig. 5B). Interestingly, the undifferentiated hES cell survival activity decreased to approximately 25% baseline from day 2 to day 7, followed by a robust rebound of cell survival activity from day 14 to day 28. This dichotomous pattern of hES cell death followed by hES cell proliferation was seen in most animals analyzed.
Post-Mortem Histology and Immunohistochemistry
To confirm our in vivo imaging data (both MR and bioluminescence), all animals were sacrificed 4 weeks after cell transplantation. Teratoma formation was observed uniformly within the left hind limbs injected with undifferentiated hES cells (supplemental online Fig. 2A) but not in the right hind limbs injected with differentiated hESC-ECs. Immunostaining for macrophages (Mac-3) revealed that in the right hind limbs injected with hESC-ECs, most of the iron particles were deposited between muscle bundles and taken up by macrophages, thus accounting for the persistent MR signals seen at 4 weeks (Fig. 6A). Immunofluorescence staining showed a similar macrophage deposition pattern, and GFP+ hESC-ECs were seldom found, but a few integrated into host microvasculatures (Fig. 6B; supplemental online Fig. 2B). In the left hind limbs injected with undifferentiated hES cells, macrophage infiltration was distributed between muscle fibers. As expected, the teratoma formation stained GFP+ uniformly throughout (Fig. 6C, 6D). The unique morphology of hES cells, which are round and small with relatively large nuclei, made them easily identifiable in the surrounding muscle.
The main aims of this study were to identify suitable molecular markers for serial monitoring of hES cell survival following transplantation and to compare the in vivo behavior of differentiated versus undifferentiated hES cells. Here, for the first time, we have compared the use of two popular cell markers, Fluc reporter gene and SPIO MR contrast agent, for labeling hES cells and their derived endothelial cells ex vivo. Under proper differentiation conditions, hES cells can be differentiated into functional endothelial cells, which express high levels of CD31 and VE-cadherin and can form tube-like structures on Matrigel assay and uptake DiI-ac-LDL. After double labeling, the fate of transplanted hES cells and hESC-ECs can be monitored by both imaging techniques. For bioluminescence imaging, a time-dependent decrease of cell signal activity was observed in the hESC-EC group, indicating significant acute donor cell loss. By contrast, engraftment of undifferentiated hES cells was followed by dramatic increases in bioluminescence signals from week 2 to week 4, which were confirmed as teratoma formation by post-mortem analysis. For MR imaging, persistent and stable signals (both GRE and OR) were seen within both hind limbs injected with undifferentiated hES cells and differentiated hESC-ECs. The negative-contrast GRE sequence of MR revealed higher anatomic resolution of the teratoma formation, and the positive-contrast OR sequence readily showed the cell location after injection. However, overall MR imaging was unable to distinguish cell viability from cell proliferation as its signals remained relatively constant over the 4-week time span.
MRI has been used for tracking mouse ES cells  and mesenchymal stem cells [21, 32, 33] in the heart and brain. However, in this study, MR imaging demonstrated stable signals in both hind limbs implanted with undifferentiated hES cells and undifferentiated hESC-ECs over 4 weeks. Thus, MR imaging was unable to specifically distinguish viable from nonviable cells or proliferating from nonproliferating cell populations. This is not surprising because, as the cells proliferate, the constant number of SPIO nanoparticles is merely divided among the daughter cells [19, 20]. Moreover, the iron from cells undergoing apoptosis or cell lyses can be internalized by resident macrophages or remain in the local tissue. Altogether, these findings suggest that MR imaging, at least in the way used in this study, cannot distinguish iron-labeled cells from free iron released upon cell death, making iron labeling a less suitable marker for tracking long-term cell survival. Finally, the divergent pattern of survival (hESC-ECs vs. hES cells) seen in our reporter gene data is also consistent with previous studies showing poor donor cell survival after adult stem cell transplantation and uncontrolled teratoma formation after introduction of undifferentiated ES cells [14, 18, 34, 35].
For bioluminescence imaging, several studies have found a close relationship between cell numbers and imaging signals [14, 18, 34, 36]. In this study, we were able to determine the kinetics of hES cell and hESC-EC survival over time. By imaging the same individual animal, we avoided the sampling biases and errors that bedevil conventional studies in which groups of animals have to be sacrificed at different time points for histology purpose . Nevertheless, at present, bioluminescence imaging still lacks adequate tomographic resolution (Fig. 5) because of attenuation of photons within tissues. To solve this problem, a combined multimodality approach (e.g., bioluminescence imaging with MR or PET) may be designed in the future as a more suitable approach to monitoring the spatial and temporal kinetics of transplanted hES donor cells in animal models in vivo.
Human ES cells are remarkable for their unlimited self-renewal and pluripotency capacity, making them highly desirable candidates for cell replacement therapy . One major risk involving the use of hES cells, however, is the possibility of cell misbehavior following transplantation. This potentially serious complication may occur if any of the transplanted undifferentiated ES cells take on teratoma formation [14, 18]. In our study, the survival kinetics of undifferentiated hES cells was different from that of predifferentiated hESC-ECs, with acute donor cell loss from day 2 to day 14 followed by a strong rebound of cell survival and proliferation from week 2 to week 4 due to subsequent teratoma formation. Thus, the ability to visualize cellular proliferation and differentiation in vivo in both predifferentiated and undifferentiated populations would be of great benefit in monitoring cellular behavior. On the other hand, endothelial cells are promising key factors for the repair of ischemic tissues and formation of new blood vessels [2, 37, 38]. Several studies have explored the endothelial potential of hES cells, mainly by demonstrating the spontaneous differentiation of EBs to vascular-like structures and isolating hESC-ECs [4, 9, 16]. In this study, hESC-ECs isolated from hEBs after 12 days of differentiation displayed characteristics similar to those of vascular endothelium and expressed typical EC markers similar to those expressed in HUVECs, such as VE-cadherin, CD31, and DiI-ac-LDL uptake. However, bioluminescence imaging data suggest that by week 4, <1.5% of the transplanted hESC-ECs are still alive. This observation conforms with other studies showing poor donor cell survival using serial histology, terminal deoxynucleotidyl transferase dUTP nick-end labeling apoptosis assay, or TaqMan Sry PCR techniques . Thus, the application of bioengineering methods or prosurvival cocktails , rather than direct stem cell injection, may prove to be a more viable approach for achieving long-term engraftment in the future .
In conclusion, although pluripotent hES cells represent a potentially unlimited source of cells for regeneration medicine, teratoma formation observed in this study and other reports [14, 18] recommends that extreme caution be exercised. Careful and precise protocols for acquiring differentiated cells are needed. To confirm the fate of hES cells in vivo, it is crucial to continue the development and further refinement of noninvasive imaging techniques. To that end, we compared the effects of labeling hES cells with reporter gene and iron particles. We showed that MR signals were persistent over a span of 4 weeks regardless of imaging undifferentiated hES cells (which leads to teratoma formation) or differentiated hESC-ECs (which leads to acute donor cell death). These data lead us to believe that reporter gene imaging is a better technique for monitoring long-term cell viability, death, and proliferation, whereas MR imaging is a better technique for high-resolution detection of cell location post-transplantation.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This work was supported in part by NIH Grants HL074883 and HL089027, a Burroughs Wellcome Foundation Career Award in Biomedical Sciences, and California Institute of Regenerative Medicine (to J.C.W.) and by a Stanford Dean's Fellowship (to Z.L.).