In vitro comparison of the biological effects of three transfection methods for magnetically labeling mouse embryonic stem cells with ferumoxides



In vivo MRI of stem cells (SCs) is an emerging application to evaluate the role of cell therapy in restoring the injured myocardium. The high spatial and temporal resolution combined with iron-oxide-based intracellular labeling techniques will provide a sensitive, noninvasive, dual imaging modality for both cells and myocardium. In order to facilitate this novel imaging approach, much effort has been directed towards developing efficient transfection methods. While techniques utilizing poly-L-lysine (PLL), protamine sulfate (PS), and electroporation (ELP) have been proposed, the fundamental biological effects of these methods on mouse embryonic SCs (mESC) have not been investigated systematically. In this study a longitudinal in vitro evaluation of cellular viability, apoptosis, proliferation, and cardiac differentiation of magnetically labeled mESC was conducted. No significant difference was seen in these biological parameters among the three transfection methods. However, cardiac differentiation was most attenuated by ELP, and iron uptake was most effective by PS. Magn Reson Med 57:1173–1179, 2007. © 2007 Wiley-Liss, Inc.

Growing evidence from preclinical and clinical studies suggests that cell therapy can improve cardiac function (1–4). Although this therapeutic option is a promising approach to repair the injured myocardium, the exact mechanism of functional restoration of the injured myocardium is unknown. While differentiation of embryonic stem cells (ESC) into functional cardiomyocytes has been attributed as one of the possible underlying mechanisms in restoring the injured myocardium, a more fundamental issue regarding cellular viability following delivery into the myocardium also needs to be addressed (5, 6). 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 aid clinical implementation of cell therapy. Imaging modalities such as single-photon emission computed tomography (SPECT), positron emission tomography (PET), magnetic resonance imaging (MRI), and optical bioluminescence imaging (BLI) have been reported to be capable of evaluating cellular engraftment parameters (4, 7, 8).

In vivo MRI of SCs is an emerging application to monitor SC engraftment. The high spatial and temporal resolution combined with robust iron-oxide-based intracellular labeling enables dual assessment of myocardial function and cellular location. Much effort has been directed towards developing efficient transfection methods to facilitate intracellular magnetic labeling with superparamagnetic iron oxide (SPIO). Although techniques utilizing poly-L-lysine (PLL), protamine sulfate (PS), and electroporation (ELP) have been proposed, the effects of these three transfection methods on engraftment parameters of ESC have not been investigated systematically (9–11). In this study a comparison of the engraftment parameters of magnetically labeled mouse embryonic SCs (mESC) using the three transfection methods was conducted. Fundamental biological determinants of cellular engraftment, including longitudinal viability, proliferation, apoptosis, and differentiation of mESC into cardiomyocytes, were compared among the three transfection techniques.


mESC Culture

The mESC were maintained in an undifferentiated state on a feeder layer of irradiated mouse fibroblasts with maintenance medium containing Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA, USA) supplemented with 15% FBS (Hyclone, Logan, UT), 0.1 mM 2-mercaptoethanol (Sigma, St. Louis, MO, USA), 0.1 mN non-essential amino acids (Invitrogen), 1% penicillin/streptomycin (Invitrogen), 2 mM L-glutamine (Invitrogen), and 1000 U/mL mouse leukemia inhibitory factor (LIF, ESGRO; Chemicon, Temecula, CA, USA) at 37°C as described previously (12). Three stable lines of mESC were employed: 1) the TL-1 mESC line derived from 129sv/J mice (courtesy of Dr. Thomas Quertermous, Stanford University, Stanford, CA), 2) luciferase-transfected mESC (mESC-luc+, courtesy of Dr. Joseph C. Wu, Stanford University, Stanford, CA), and 3) Nkx2.5 (cardiomyocyte differentiation specific promoter)-green fluorescent protein (GFP) transfected mESC (mESC-Nkx2.5-GFP+, courtesy of Dr. Takayuki Morisaki, National Cardiovascular Center Research Institute, Osaka, Japan) were used. All experiments, assays, and measurements described below utilizing the mESC were performed at least in triplicate.

Cell Labeling

To label the mESC we used a commercially available ferumoxides suspension (Feridex IV; Berlex Laboratories, Wayne, NJ, USA) that contains particles approximately 80–150 nm in size and has a total iron content of 11.2 mg Fe/mL.


A solution containing PLL (1.5 μg/mL, molecular weight = 388 kDa; Sigma, St. Louis, MO, USA) and ferumoxides was added to the mESC maintenance medium in a rotating mixer for 1 hr. The labeling solution was then added in a 1:1 volume ratio to the mESC culture and incubated for 12–24 hr. After incubation the cells were washed twice with sterile phosphate-buffered saline (PBS) (9).


Clinical-grade PS (American Pharmaceuticals Partner, Schaumburg, IL, USA) was prepared as a stock solution of 1 mg/mL in distilled water. Ferumoxides was put into a tube containing serum-free RPMI 1640 medium (Invitrogen) containing 25 mM HEPES and L-glutamine. PS was added to that solution at a concentration of 12 μg/mL. The solution containing ferumoxides and PS was mixed for five to 10 min. An equal volume of labeling solution was added to the existing medium in the mESC culture and incubated for 12–24 hr. After labeling, the cells were washed twice with PBS, with the last wash containing heparin (10 U/mL) to dissolve the extracellular ferumoxides-PS complex (10, 13).


Following suspension in PBS at a density of 1–5 × 106 cells/mL in sterile 0.4-mm-gap ELP cuvettes (Bio-Rad, Hercules, CA, USA), the mESC were electroporated with the Gene Pulser II System (Bio-Rad) using an optimized condition (voltage: 800V, capacitance: 3 μF). Following ELP with ferumoxides, mESC were left in the cuvette holder for one min, left on ice for five min, and washed twice with PBS (11). Then the number of viable cells was determined by trypan blue exclusion and only the viable cells were subjected to viability, proliferation, apoptosis, and ROS studies.

In Vitro Cellular MRI

In vitro cellular MRI (N = 5 cell preparations) was performed at 1.5T (Signa; GE Medical Systems, Milwaukee, WI, USA) using a five-inch receive-only surface coil. The cell suspension was prepared in 96-well plates (PCR plate; E & K Scientific, Santa Clara, CA, USA) and MRI was performed using a gradient-echo sequence (TR = 100 msec, TE = 10 msec, flip angle = 30°, FOV = 12 × 12 cm, slice thickness = 1 mm, and matrix = 256 × 256). The area of signal dephasing was measured by using Image J (NIH, Bethesda, MD, USA).

Cellular Viability, Proliferation, and Apoptosis

MTT Assay

The viability and proliferation of labeled mESC were evaluated by a 3-[4,5-dimethylthiazol-2-yl]-2, 5-diphenyl tetrazolium bromide (MTT) assay (Sigma, St. Louis, MO, USA) at days 1, 4, 7, and 14 after cell labeling (PLL, PS; 50μg Fe/mL, ELP; 2 mg Fe/mL). The MTT assay is a means of measuring the activity of living cells via mitochondrial dehydrogenase. The mESC (N = 3 cell preparations) were initially seeded in 48-well plates at 5 × 103 cells per well. MTT was added in an amount equal to 10% of the culture medium volume. After incubation for three hr at 37°C, MTT solubilization solution equal to the original culture medium volume was added and the resulting formazan crystals were dissolved. The absorbance of the formazan product was measured spectrophotometrically at 570 nm as the reference wavelength according to the manufacturer's recommendations.

Optical BLI

To evaluate the longitudinal cellular viability and proliferation, optical BLI was acquired at days 1, 4, 7, and 10 following cell labeling with multiple doses of ferumoxides (PLL, PS; 50, 100, and 200 μg Fe/mL; ELP; 2, 4, and 8 mg Fe/mL) to assess the activity of luciferase in mESC-luc+ cells. BLI was performed using the charged coupled device camera (Xenogen) (14). Bioluminescence was quantified in units of photons · s−1 · cm2−1 · sr−1. The mESC (N = 3 cell preparations) were initially seeded in 24-well plates at 2 × 104 cells per well. D-luciferin was directly administered to the culture medium at a dose of 30 μg/mL (Xenogen Corp.). The cells were imaged for 30 min using 1-min acquisition scans.

Apoptosis Detection by Annexin V

At days 1, 4, 7, and 14 following cell labeling (PLL, PS; 50 μg Fe/mL; ELP; 2 mg Fe/mL), the percentages of dead cells and cells undergoing apoptosis were determined by means of the Annexin V-Biotin apoptosis detection kit (Calbiochem, San Diego, CA, USA) using fluorescent-activated cell sorter (FACS) (Becton Dickinson, San Jose, CA, USA). After the cells were trypsinized and collected, the cell suspension concentration was adjusted to approximately 1 × 106 cells/mL. Then, according to the manufacturers' protocols, 10 μL media binding reagent and 1.25 μL Annexin V-Biotin were added to the cell suspension (N = 4 cell preparations) and incubated for 15 min at room temperature (18–24°C) in the dark. Then 10 μL propidium iodide were added and the cells were analyzed by FACS. A total of 30000 cells/analysis were measured at the FITC detector and the results were presented as the mean geometric fluorescence.

Reactive Oxygen Species (ROS) Measurements

The intracellular formation of ROS was detected by using a CM-H2DCFDA fluorescent probe (Molecular Probes, Eugene, OR, USA), a nonfluorescent indicator that forms fluorescent esters when it reacts with ROS inside cells. Labeled and unlabeled cells (N = 3 cell preparations) were grown for ROS analysis (PLL, PS; 50μg Fe/mL; ELP; 2 mg Fe/mL). At days 1, 4, 7, and 14, 1 × 106 cells were suspended in PBS with a final concentration of CM-H2DCFDA of 10 μM and incubated for 1 hr at 37°C in a 5% CO2 atmosphere. Thereafter the cells were washed twice and resuspended in PBS. An increase in fluorescence was detected by FACS. A total of 30000 cells/analysis was measured at the FL2 detector and the results were presented as the mean geometric fluorescence.

Cellular Cardiac Differentiation

Contractile Embryoid Bodies (EBs)

The differentiation capability of the labeled and unlabeled mESC using mESC-TL-1 and mESC-Nkx2.5-GFP+ was assessed. Following cell labeling (PLL, PS; 50 μg Fe/mL; ELP; 2 mg Fe/mL), the hanging drop method was employed to form EBs, a 3D cellular aggregate to facilitate cardiomyocyte differentiation from undifferentiated mESC. The EBs were formed from the mESC (N = 3 cell preparations generating EBs) for 2 days (days 0–2) with mESC differentiation medium containing Iscove's modified Dulbecco's medium (Invitrogen) supplemented with 15% FBS (Hyclone), 2 mM L-glutamine (Invitrogen), 1% insulin-transferrin-selenium (Invitrogen), and 0.1 mM 2-mercaptoethanol (Sigma, St. Louis, MO, USA). The EBs were maintained in suspension for three days (days 2–5) before being plated onto gelatin-coated plates (15). At day 10 the percentage of EBs that differentiated into a cardiac phenotype, as defined by both spontaneous contractility observed under the microscope and Nkx2.5-GFP-positive signal detected by FACS (Becton Dickinson, San Jose, CA, USA), was determined. A total of 30000 cells/analysis were measured at the FL3 detector and the results were presented as the mean geometric fluorescence.

Taqman RT-PCR

In order to assess the levels of cardiac phenotype gene expression, total RNA was isolated using an RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA) at day 14. For each RNA sample, cDNA synthesis was performed using a iScript™ cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA) following the manufacturer's two-step protocol. Quantitative analysis of cardiomyocyte-specific α-cardiac myosin heavy chain (α-MHC) and cardiac myosin light chain (MLC2v) genes was performed by real-time quantitative PCR using the ABI Prism 7700 Sequence Detection System (PE Applied Biosystems) with TaqMan® chemistries and probes. The TaqMan® probes and primers for target genes were assay-on-demand gene expression products (PE Applied Biosystems). The following validated PCR primers and TaqMan® MGB probes (FAM-labeled) were used: mouse αMHC (assay ID: Mm00440354_m1), mouse MLC2v (assay ID: Mm004400384_m1), and 18S rRNA (assay ID: 4319413E) as endogenous control (PE Applied Biosystems). Three measurements per sample were performed in each of three independent experiments. Results were analyzed with the ABI Sequence Detector software version 1.7 (PE Applied Biosystems). For relative quantification of target gene expression, the standard curve method was applied.

Cellular Iron Content

The iron content of the labeled cells (N = 5 cell preparations) was assessed by Prussian blue (PB) colorimetric assay. Then 100 μL of a 5% solution of potassium ferrocyanide (prepared in demineralized water) were added to 100 μL of iron solution (after mineralization by 5N HCl). The optical density was read at 650 nm after a 15-min incubation. We initially determined the calibration curve using different concentrations of ferumoxides ranging between 0 and 220 μg Fe/mL. A linear correlation between the optical density and the iron concentration was obtained (absorbance = 0.870 for a 100-μg Fe/mL concentration).

PB Staining

After the cells were labeled (PLL, PS: 50 μg Fe/mL; ELP: 2 mg Fe/mL), they were fixed with 4% paraformaldehyde (N = 5 cell preparations). Then they were washed with distilled water and incubated for 20–30 min with a working solution containing 5% potassium ferrocyanide (Perl reagent for PB staining) in 5% hydrochloric acid, washed again, and counterstained with nuclear fast red.

Statistical Analysis

All data are expressed as the mean ± standard deviation (SD). To identify significant differences in the viability, area of signal dephasing, and iron content of the cells, a one-way analysis of variance (ANOVA) was applied. A P-value < 0.05 was considered statistically significant.


In Vitro Cellular MRI, Histology, and Iron Content

The PS-labeled cells demonstrated the largest area of dephasing signal (PLL: 22.5 ± 0.2 mm2; PS: 37.5 ± 0.6 mm2; ELP: 28.1 ± 1.6 mm2; P < 0.0001). Representative in vitro cellular MRI of labeled cells is shown in Fig. 1. Similarly, intracellular iron content observed with the PB colorimetric assay demonstrated that the amount of cellular iron content of PS was significantly higher compared to other groups (P < 0.05), as shown in Table 1. Representative PB-stained labeled cells are shown in Fig. 2.

Figure 1.

In vitro cellular MRI of labeled cells by three transfection methods. The cell suspensions in 96-well plates each contained 1 × 106 labeled mESC. The ferumoxides concentrations used for cell labeling in this figure are as follows: PLL: 50 μg Fe/mL; PS: 50 μg Fe/mL; and ELP: 2 mg Fe/mL. PS demonstrated the largest area of signal dephasing (P < 0.05). Data are derived from five samples of each group.

Table 1. Iron Content in Labeled mESC by Three Transfection Methods (g/cell)*
Fe added (μg Fe/mL)PLLPSELPP
  • *

    The amount of cellular iron uptake of PS was significantly higher than other groups. All data are expressed as means ± 1 SD. Data are derived from five samples.

  • ns = not significant.

25 μg Fe/mL (ELP; 1 mg Fe/mL)2.41 ± 0.732.86 ± 0.482.40 ± 0.67ns
50 μg Fe/mL (ELP; 2 mg Fe/mL)8.94 ± 1.1211.91 ± 1.0110.44 ± 1.160.004
100 μg Fe/mL (ELP; 4 mg Fe/mL)12.62 ± 1.3520.7 ± 1.3114.49 ± 1.42<0.001
Figure 2.

PB-stained labeled cells: (a) unlabeled, (b) PLL-labeled, (c) PS-labeled, and (d) ELP-labeled mESC. The magnification of all slides is ×40.

Cellular Viability, Proliferation, and Apoptosis

Cellular Viability—Optical BLI

Cellular viability determined by the activity of mESC-luc+ was measured on days 1, 4, 7, and 10 following cell labeling. A viability-specific signal increase was demonstrated over time. Among the three transfection methods, no difference was observed in the cellular signal of viability. No significant difference was observed between unlabeled and labeled cells at each time point. Representative serial images and data are shown in Fig. 3.

Figure 3.

Cellular viability by optical BLI. Serial BLI of the mESC-luc+ activity by (a) PLL, (b) PS, and (c) ELP. The bar on the right indicates the bioluminescent scales and analysis of the optical imaging data: (d) PLL, (e) PS, and (f) ELP. No significant difference was seen between the corresponding unlabeled and labeled cells at each time point. Data are derived from three samples of each group.

Cellular Proliferation—MTT Assay

The proliferative capability of the labeled and unlabeled cells was measured on days 1, 4, 7, and 14 using the MTT assay. No significant difference was observed among the three transfection methods or between labeled and unlabeled cells at each time point, as shown in Table 2. However, we observed that the number of viable cells immediately after ELP labeling decreased (62.3% ± 5.5% of the number of viable cells before ELP labeling).

Table 2. Viability of Labeled mESC Measured at Different Time Points Using MTT, Apoptosis, and ROS Assays*
 Day 1Day 4Day 7Day 14
  • *

    All data are expressed as means ± 1 SD. Data are derived from 3-4 samples (MTT; N = 3, annexin V; N = 4, ROS; N = 3). Data are mean percentages of the average viability values for the corresponding control cells.

PLL (%)    
 MTT96.3 ± 2.1107.4 ± 8.0111.6 ± 6.0113.1 ± 6.4
 Annexin V assay109.2 ± 9.599.1 ± 6.1102.5 ± 8.4112.8 ± 15.5
 ROS104.8 ± 11.1100.4 ± 4.8106.4 ± 8.9107.4 ± 5.7
PS (%)    
 MTT89.7 ± 3.393.2 ± 5.2109.7 ± 7.7119.3 ± 1.6
 Annexin V assay106.5 ± 16.9106.8 ± 5.6113.0 ± 9.3107.2 ± 9.8
 ROS102.1 ± 16.6108.4 ± 7.798.0 ± 1.9102.5 ± 5.1
ELP (%)    
 MTT119.0 ± 2.6123.5 ± 2.1105.2 ± 2.4102.9 ± 1.4
 Annexin V assay102.6 ± 15.1103.1 ± 9.1110.3 ± 6.3115.6 ± 6.4
 ROS106.6 ± 9.5104.5 ± 9.1111.5 ± 8.4110.8 ± 7.9

Apoptosis—Annexin V and ROS FACS Analysis

The proportions of apoptotic labeled and unlabeled cells were compared on days 1, 4, 7, and 14 using FACS analysis of Annexin V and ROS. No significant difference among the three transfection methods or between labeled and unlabeled cells was observed at each time point, as shown in Table 2.

Cellular Cardiac Differentiation

The capacity of mESC to differentiate into cardiomyocytes was determined by the percentage of beating EBs. No significant difference in cardiac differentiation capacity was demonstrated between the PLL and PS transfection groups when compared to the unlabeled mESC. However, the ELP-transfected group demonstrated a lower percentage of beating EBs (unlabeled control 84.9% (N = 90/106; beating EB/total EB), PLL 79.4% (N = 77/97), PS 80.8% (N = 80/99), and ELP 68.9% (N = 71/103), P < 0.05). The gene expression levels of α-MHC and MLC2v were significantly attenuated in ELP-labeled mESC, as shown in Fig. 4. Similar findings were also demonstrated in the FACS analysis of % Nkx 2.5 GFP+ cells (unlabeled control 1.45%, PLL 1.05%, PS 1.15%, and ELP 0.53%) as shown in Fig. 5.

Figure 4.

Cardiac gene expression of labeled and unlabeled mESC. The cardiac gene expression levels of (a) αMHC and (b) MLC2v in labeled and unlabeled mESC after induction of differentiation were determined by RT-PCR. Both gene expression in ELP-labeled mESC were lower than those in other groups. *: P = 0.132, §: P = 0.01. Data are derived from three samples of each group.

Figure 5.

Cardiac differentiation by Nkx2.5-GFP+ mESC. Analysis of the percentage of Nkx2.5-GFP+ cells in samples of (a) unlabeled, (b) PLL-labeled, (c) PS-labeled, and (d) ELP-labeled mESC using FACS is shown. The region and the number in the right lower quadrant indicate GFP-positive cells and the percentage of GFP-positive cells in the gated cell population, respectively.


In this study we systematically compared the engraftment parameters of magnetically labeled mESC using three transfection methods. Ferumoxides are highly negatively charged SPIO nanoparticles that do not adhere to the cell membrane without modification of the surface charges. Polycationic transfection agents (TAs), such as PLL and PS, bind to the dextran coat through electrostatic interactions, thereby modifying the ferumoxides' distribution of positive and negative surface charges that can adhere to the cell membrane (10). ELP is known to cause reversible formation of hydrophilic pores that provide direct communication between the extracellular compartment and the cytoplasm (11).

The longitudinal cellular viability and proliferative capability did not show any difference among the three methods, consistent with previous reports (9–11, 13). However, the number of viable cells decreased immediately after ELP labeling. Furthermore, PS labeling technique demonstrated the largest MRI signal dephasing area, consistent with the highest intracellular iron uptake confirmed by PB colorimetric assay.

In terms of cardiac differentiation capability, the spontaneous contractility of EBs and gene expression of αMHC, MLC2v, and NKx2.5 provided both biological and lineage-specific detection of cardiomyocyte differentiation. This study suggests that ELP-labeled cells demonstrate the least level of cardiac differentiation at both cellular and molecular levels. Inhibition of chondrogenesis in the magnetically labeled human mesenchymal SCs using PLL has been reported (16). However, labeling of hematopoietic SCs using PS did not interfere with the differentiation of hematopoietic SCs into colony-forming units (CFU) and dendritic cells. Similarly, PS labeling of mesenchymal SCs (MSC) enabled proper differentiation of MSC into adipogenic, osteogenic, and chondrogenic lineages (13). In our study there was no significant difference between the PLL- and PS-transfected groups in terms of differentiation potential, while the differentiation of the labeled mESC employing ELP demonstrated a significant decrease.

Currently, there is a great interest in the longitudinal tracking of SC migration using MRI (4, 17, 18). Although these methods are very simple techniques, both PLL and PS require overnight incubation with ferumoxides. On the other hand, ELP provides rapid magnetic labeling of SCs, which is ideal for the clinical setting. However, this method requires a large amount of iron for cell labeling and is known to decrease the number of viable cells due to electric shock. Assessment of SC engraftment will mandate the tracking of smaller quantities of viable cells, and thus require higher intracellular iron uptake. Our study suggests that PS will provide optimal conditions for in vivo MRI of SC engraftment in the myocardium. Furthermore, PS has the additional advantage of being a clinical-grade reagent approved for human use by the FDA.


Magnetic labeling of mESC using the three transfection methods is safe and effective. Although there was no significant difference in the viability and proliferative capability of mESC among the three methods, the cardiac differentiation capability of mESC transfected by ELP was the most attenuated, and iron uptake by PS was the most efficient.