|Home | About | Journals | Submit | Contact Us | Français|
Development of non-invasive and accurate methods to track cell fate following delivery will greatly expedite transition of embryonic stem (ES) cell therapy to the clinic. Here we describe a protocol for the in vivo monitoring of stem cell survival, proliferation, and migration using reporter genes. We established stable ES cell lines constitutively expressing double fusion (DF; enhanced green fluorescent protein and firefly luciferase) or triple fusion (TF; monomeric red fluorescent protein, firefly luciferase, and herpes simplex virus thymidine kinase) reporter genes using lentiviral transduction. We used fluorescence activated cell sorting to purify these populations in vitro, bioluminescence imaging and positron emission tomography imaging to track them in vivo, and fluorescence immunostaining to confirm the results ex vivo. Unlike other methods of cell tracking such as iron particle and radionuclide labeling, reporter genes are inherited genetically and can be used to monitor cell proliferation and survival for the lifetime of transplanted cells and their progeny.
Embryonic stem (ES) cells are capable of differentiation into any somatic cell type of the human body and have the potential for unlimited self renewal 1. As a result, these cells have been regarded as a leading candidate source for donor cells in regenerative medicine. Before ES cells can be safely applied clinically, however, it is important to understand the in vivo behavior of ES cells and their derivatives. Conventional histology and reporter genes such as green fluorescent protein (GFP) and β-galactosidase (LacZ) do not allow for longitudinal imaging of cells that have been injected into animals because these methods require animal sacrifice and at best only provide a “snapshot” of the biological fate of transplanted cells 2, 3. Recent advances in the field of molecular imaging have made it possible to non-invasively track transplanted cells over time 4. These modalities include physically attaching labels to cells as in the case of iron particles, radionuclide probes, and quantum dots, or introducing reporter genes into cell lines to obtain the cell-mediated generation of reporter probes 5.
Traditional methods of imaging cell delivery in vivo have typically relied upon physical cell labeling as these modalities provide a straightforward approach to visualizing transplanted cells in living subjects 6, 7. Physical cell labeling is completed before cell administration and can be accomplished with superparamagnetic iron oxide (SPIO) particles for magnetic resonance imaging (MRI) 8, 9 nanoparticle labeling for fluorescent imaging 10, 11, or radionuclide labeling for single positron emission computed tomography (SPECT) or positron emission tomography (PET) 12. Physical cell labeling allows high spatial resolution (MRI) and high sensitivity (SPECT or PET) imaging and is best used to track the in vivo localization of cells in the hours to days following delivery. However, a common drawback these methods share is their reliance on physical labels. SPIO and radionuclide probes are diluted with cell division and are not capable of tracking cell proliferation, especially when cells misbehave as in the case of ES cell-derived teratoma formation 13. SPIO agents further suffer from the unique problem of being taken up by macrophages after donor cell death (which may continue to produce signal even after cell death), and hence cannot be used to accurately monitor long-term cell survival and behavior 14. By comparison, SPECT or PET tracers lose signal due to radioisotope decay. A typical PET radioisotope such as F-18 has a half-life of only 110 minutes and can only be used to image cells in the hours immediately following cell delivery 15. Other isotopes such as 99mTechnetium and 111Indium have half-lives on the order of 6 hours and 2.8 days, respectively, and can be used to image cells for several hours to close to a week 16, 17. However, in these studies, because cells are exposed to longer periods of radioactivity, impairment of cellular proliferation and differentiation may be an issue.
In contrast with physical cell labeling modalities, reporter gene imaging is well suited for longitudinal imaging of cell survival. In this type of imaging, a gene coding for the synthesis of a detectable protein is introduced into a target cell line or tissue via viral or non-viral vectors. Examples of commonly used reporter genes include firefly luciferase (Fluc) and herpes simplex virus thymidine kinase (HSVtk), which can be detected by bioluminescence imaging (BLI) and PET, respectively 5. Reporter genes can be inserted after a constitutive promoter such as ubiquitin, or after a tissue specific promoter such as myosin light chain in the case of cardiomyocytes 18. BLI has a sensitivity on the order of 10−15 to 10−17 mol L−1, whereas the sensitivity of PET is 10−10 to 10−11 mol L−1, and MRI has a sensitivity of only 10−3 to 10−5 mol L−1 5. Importantly, because active transcription of the reporter gene is a prerequisite for synthesis of the reporter protein, only cells that are alive yield positive imaging signals. Several studies have shown that BLI signals correlate robustly with cell numbers both in vitro and in vivo 19, 20. Hence, changes in signals following cell administration can be used as indicators of cell engraftment or cell death. In addition, because the reporter gene integrates into the host cell’s chromosome following stable transfection or transduction, the reporter gene is passed on from the mother cell to daughter cell. Genetic inheritance of the reporter gene thus permits monitoring of donor cell proliferation (e.g., ES cell-derived teratoma formation). Finally, genomic and proteomic studies have shown that reporter genes do not significantly affect ES cell viability, proliferation, or differentiation 21, 22.
Using BLI and PET reporter gene imaging, our laboratory has successfully monitored the survival, proliferation, and migration of transplanted ES cells 19, 23, 24 and their derivatives, such as cardiomyocytes 25 and endothelial cells 13, 26 over a prolonged period without necessitating animal sacrifice. We have also monitored immunogenic response against ES cell engraftment in syngeneic, allogeneic, and xenogenic transplantation models 23, 24. As reporter genes can be inserted after any promoter, this approach can also be used to monitor expression of target genes in developmental pathways and disease models. Our group has successfully applied BLI in this fashion to investigate patterns of STAT3 expression in embryoid body formation 27 and to track plasmid-mediated transgene expression for short hairpin RNA interference therapy in C2C12 myoblasts 28.
For delivery of the reporter gene, our group has employed both a double fusion (DF) construct containing enhanced green fluorescent protein (eGFP) and firefly luciferase (Fluc) and a triple fusion (TF) construct containing monomeric red fluorescent protein (mRFP), Fluc, and HSVtk (Figure. 1a). When stably integrated into the genome of the cells, these constructs permit the longitudinal tracking of transfected cells using a multimodality imaging approach. Specifically, eGFP and mRFP reporter genes facilitate fluorescence microscopy and FACS sorting of GFP/RFP positive cells, whereas the Fluc and HSVtk reporter genes allow for cell monitoring via BLI and PET, respectively. Interaction of Fluc with its substrate D-luciferin produces low intensity light (2–3 eV) that is detected by an ultrasensitive cooled charge-coupled device (CCD) camera for cell localization (Figure. 1b). The reporter protein HSVtk phosphorylates its substrate, the PET reporter probe 9-4-[18F]fluoro-3-(hydroxylmethylbutyl) guanine ([18F]-FHBG), to produce high-energy photons (511 keV). These photons are then captured by the PET camera for cell localization (Figure. 1b) in a fashion identical to radiotracer based PET imaging. The advantage of reporter gene PET imaging over radiotracer labeled PET imaging is that constitutive expression of the reporter protein HSVtk allows for longitudinal tracking of cell survival and localization without the constraint of label decay.
The choice of whether to use physical cell labeling or reporter gene modalities for in vivo imaging of transplanted cells depends on the subject of investigation, timeline of study, equipment available at a given institution, and evaluation of the advantages and disadvantages of each technique (Figure. 2). However, it is important to note that the use of separate imaging modalities such as MRI, PET, SPECT, or BLI is not mutually exclusive. Multimodality imaging approaches may minimize the potential drawbacks of using each imaging modality alone and a tailored combination of 2 or more techniques may be the best approach for a given experiment.
Reporter gene imaging suffers from several drawbacks. First, derivation of stable DF and TF reporter gene positive cell lines typically takes 2–4 weeks, whereas preparation for iron particle or radioactive probe labeling can be completed within hours. Second, the spatial resolution of common reporter gene modalities such as PET or BLI is close to 1 mm3 or 3 mm3, respectively, whereas MRI has a spatial resolution of 25–100 μm. BLI can only provide a general anatomical location of where the cells have engrafted because deep tissue attenuates light. Because photons generated by the interaction of Fluc and D-luciferin can only penetrate 1–2 cm of tissue, this technology is primarily limited to small rodent models at the present time. By comparison, PET reporter gene imaging does not have these constraints as photons emitted from the phosphorylation of [18F]-FHBG are relatively high energy 5. The use of the HSVtk reporter gene and [18F]-FHBG reporter probes have recently been demonstrated to track mesenchymal stem cell fate in a porcine model 29 and cytolytic T cells in humans 30. Third, the introduction of reporter genes has the potential to alter cellular genome and phenotype.
This protocol details how reporter gene imaging may be used to monitor the engraftment, survival, and proliferation of transplanted ES cells. We hope it may be used in conjunction with other imaging technologies such as MRI and radiotracer based techniques to answer a wide range of biological questions.
The DF and TF lentiviral vectors can be constructed using the conventional molecular cloning techniques based on restriction enzymes, For general information of cloning techniques, please see ref. 31. We have used a constitutive ubiquitin promoter to replace the cytolomegavirus (CMV) promoter to reduce potential transgene silencing after extended cell culture 32. We use PEG-it Lentivirus Concentration Solution to precipitate and concentrate the lentiviral particles, which results in lower toxicity to the transduced ES cells compare to viruses concentrated by ultracentrifugation. The choice of cell type depends on the focus of specific studies and whether the cells are suitable for stable cell line generation. Our lab has generated several DF and TF mouse and human ES cell lines such as the mouse D3 and human H7, H9, and HES2 lines using the protocol described below. In our experience, mouse ES cells are easier to form undifferentiated colonies on the mouse embryonic fibroblast (MEF) feeder layer from FACS sorted single cells compare to human ES cells, and thus easier to be isolated from surrounding MEFs. We therefore typically transduce mouse ES cells on feeder layers, whereas human ES cells are transduced under feeder-free conditions to eliminate any contamination from GFP/RFP positive MEFs. Using a multiplicity of infection (MOI) of 10 usually gives rise to the highest transduction efficiency for ES cells. The optimal MOI for other type of cells should be experimentally determined. When transducing ES cells, colony sizes of 200–400 cells/colony usually yield the highest transduction efficiency. If the ES cell colonies are too large, cells that are compacted in the center will not be readily transduced. Successful transduction can be verified by observing the GFP/RFP positive cells under a fluorescence microscope and the approximate transduction efficiency can be calculated by counting GFP/RFP positive and negative cells using flow cytometry. Generally a transduction efficiency of 30–40% is sufficient for FACS and subsequent subculture. Slightly lower transduction efficiency is tolerable but requires more starting cells for sorting. We recommend cryopreserving some of the ES cells derived from the first round of FACS sorting and expansion for any future experiments. A second round of FACS sorting for GFP/RFP following the initial subculture is needed to further isolate a highly purified population of cells that are positive for GFP/RFP. Following each round of sorting, some ES cell colonies may differentiate. Therefore, isolated mouse and human ES cells should be seeded on a MEF feeder layer and only colonies that are characterized by typical ES cell morphologies 33, 34 (refractive appearance, defined boundaries, and high nuclear-to-cytoplasm ratio within individual cells) should be subcultured. The highly purified DF or TF ES cells that expanded from the second round of sorting are ready for transplantation into experimental animals.
After establishing the DF/TF stable cell line, BLI of reporter gene positive cells growing on cell culture plates is helpful to determine whether the Fluc transgene is functional. This assay will also confirm that BLI signals (plotted in units of maximum photons per second per centimeter square per steridin (photons/s/cm2/sr)) correlate with cell numbers (Figure 3, a and b). For transplanting ES cells into animals, the DF/TF ES cells are usually trypsinized and collected as a single cell suspension. For in vivo imaging, the minimum number of cells detectable by BLI is approximately 100–500 cells 35 and 1,000 cells/mm3 for small animal PET imaging 36. However, this detection threshold number can vary depending on the robustness of the promoter or enhancer element used to drive the reporter gene, the specific cell type, and the amount of reporter probe used in each study. Typically our laboratory has used anywhere from thousands to millions of cells for purposes of injection and longitudinal monitoring of cell survival. While there is no ‘set’ amount of luciferin or [18F]-FHBG to administer to animals for BLI or PET imaging, respectively, delivery of inadequate reporter probe may compromise visualization of signal. Our laboratory has successfully imaged cell survival using a D-luciferin concentration of 375 mg/kg animal body weight for BLI and an [18F]-FHBG activity of ~200 μCi for microPET 37. Animals that receive transplantations of non-transduced cells can be used as negative controls to determine the background BLI and PET signals. Afterwards, cells can be imaged non-invasively at any time-point following transplantation. In the case of ES cells, acute cell death following transplantation will be observed for the first one or two weeks, after which the remaining cells will proliferate and eventually cause teratoma formation. To monitor this bimodal process, our laboratory has traditionally imaged animals receiving transplantation of ES cell or ES cell derivatives at day 0, day 2, day 7, day 10, day 14, and weekly thereafter for up to one year using BLI 25. By comparison, we have typically taken microPET images at weekly intervals due to the high cost and temporal constraints of [18F]-FHBG production and radionuclide decay. In both BLI and microPET, increases in signal indicative of cell proliferation and teratoma formation can be observed as early as the second or third weeks following ES cell transplantation. The choice of anatomical location to inject cells depends on the focus of investigation. The majority of our experiments have revolved around myocardial injection (Figure. 3) due to our group’s focus on cardiovascular disease 19, 25. However, we have also used other locations such as the gastrocnemius muscle (Figure. 2) 13, 23, 24, kidney capsule, and subcutaneous injection 23, 38. Certain locations such as the kidney and heart require technical expertise for injection. For general tracking of cell survival in vivo, we recommend delivering cells to an easily accessible location such as leg muscle or subcutaneous regions of the dorsal flank.
Operation of BLI and PET reporter gene imaging requires the use of equipment such as a Xenogen In Vivo Imaging System and a MicroPET scanner (see equipment set up below). For operation of these machines, manufacturer’s instruction should be followed. For BLI, users must have access to the software program Living Image (Caliper Life Sciences) and be able to administer D-luciferin via intra-peritoneal injection to study animals. The use of the PET scanner requires training in radiation safety as [18F]-FHBG is a radioactive probe. Typically, [18F]-FHBG is administered via tail vein injection and imaged at 60 minutes post injection to allow for adequate biodistribution and background clearance of the tracer. Animals that have not received cell injection may be used as controls to determine background signal. A software program such as ASI Pro (Concorde Microsystems) can be used to acquire images and calculate tracer uptake in units of injected dose per gram of heart (%ID/g).
DMEM high glucose with L-glutamine, 10% (vol/vol) FBS, 0.5%(vol/vol) penicillin/streptomycin. Store at 4 °C for up to one month.
Knockout DMEM, 15% (vol/vol) FBS defined, 0.1 mM nonessential amino acids, 0.1 mM 2-mercaptoethanol, 1 × 103 units/ml LIF. Store at 4 °C for up to two weeks or at −20 °C in aliquots for up to three months.
The Xenogen optical imaging system consists of a light-tight box with a mounted cooled charge couple device (CCD) camera (IVIS). The system is fully calibrated using a standard “hockey-puck” with scintillation cocktails with four small point sources of light. Standard software routines are provided with the system.
The small animal PET scanners should be set up following the manufacturer’s instructions. The two scanners we use are manufactured by GE healthcare (eXplore Vista) and Concorde Microsystems (microPET Rodent R4).
Trouble shooting advice can be found in Table 1.
Stably transduced DF and TF cells do not significantly differ from untransduced counterparts in terms of cell viability and proliferation. The integration of the DF/TF construct has been shown not to impact differentiation of ES cells 21, 22. Our laboratory has routinely differentiated these into EBs, cardiomyocytes, and endothelial cells (see ref. 8, 12, 18, and 19). Following establishment of stable DF or TF cell lines, imaging of cell survival following transplantation can be maintained for the duration of cell survival. As previously stated, estimates for the minimum number of cells detectable by BLI is approximately 500 cells 35 and 1,000 cells/mm3 for small animal PET imaging 36. However, the detection sensitivity will likely vary depending on the robustness of reporter gene expression, amount of reporter probes administered, location of transplanted cells (for BLI), and degree of cell survival. Typically following injection of undifferentiated ES cells, an acute period of cell death will be observed for the first week, reflected by a decrease in BLI signal. After that, an increase in signal will be observed due to teratoma formation (Figure 2c and upper panel of Figure 3c). In our experience, the number of cells that required for teratoma formation vary depending on the site of administration and whether the host is immunocompetent or immunodeficient. For immunodeficient host, we have been able to form tumors with as few as 500–1000 cells 38. Generally the more the blood support from the transplantation site the less cells needed for teratoma formation. By contrast, injection of differentiated hES cell-derived endothelial cells (Figure 2c) and cardiomyocytes (Figure 3c, lower panel) lead to significant cell death within the first 4 weeks, which is reflected by gradual decrease in signals (Figure 2c and Figure 3c). In the case of ES-derived cardiomyocytes, we have observed stable engraftments after transplantation out to greater than 6 months 25. For small animal PET imaging, DF cells are not appropriate for this modality because they do not carry the HSVtk reporter gene that allows uptake of the [18F]-FHBG reporter probe (Figure 4a). Only TF stable cell lines that carry the HSVtk reporter gene should be used. After administration of [18F]-FHBG into animals, the tracer will distribute to most of the tissues and clear gradually within an hour or so. Due to the natural excretion route, background PET activity will generally be present in the liver and bladder region (Figure 4a). For transplanted ES cells, a significant increase in PET signals will be observed after one week of transplantation.
We have also compared MRI and BLI modalities in tracking the cell fate of transplanted DF positive hES cells and hES cell-derived endothelial cells (hESC-ECs). Although MRI can detect the teratoma formation of DF hES cells labeled with iron particles at day 28 after transplantation (Figure 2a), it is not appropriate for imaging the process of cell proliferation over time (Figure 2b). Macrophages loaded with iron particles could be found in between muscle bundles that are close to the site of injection, which explains why MRI signals were relatively constant over the 4 week period post transplantation (Figure 2e). Because of the poor survival of hESC-ECs, there were no transplanted GFP+ cells detected nearby the macrophages that have engulfed the iron particles (Figure 2f). In contrast, BLI reporter gene imaging of the same animal exhibited a bi-modal curve (Figure 2d). The right hind limb (hESC-ECs) showed significant BLI activity at day 2, which decreased progressively over the following 4 weeks, indicating acute donor cell death. The left hind limb (undifferentiated hES cells) showed initial decrease in BLI signals at day 7, which increased dramatically during week 2 and week 4, indicating teratoma formation (Figure 2c).
In summary, following establishment of stable DF or TF cell lines, imaging of ES cell survival following transplantation is possible for the duration of cell survival. Incorporation of these imaging modalities described here will allow investigators us to study biologically relevant questions such as tumorigenicity, immunogenicity, and differentiation.
We thank funding support from AHA 0970394N, NIH HL089027, and NIH DP2OD004437 (JCW).
Competing financial interests
The authors declare that they have no competing financial interests.
Author contributionsN.S. and A.L. contributed equally to the preparation of this manuscript. J.C.W provided advices and proofread the manuscript.