Ferritin Overexpression for Noninvasive Magnetic Resonance Imaging–Based Tracking of Stem Cells Transplanted into the Heart

Abstract

An unmet need in cardiac cell therapy is a noninvasive imaging technique capable of tracking changes in graft size over time and monitoring cell dynamics such as replication and death, factors to which commonly used superparamagnetic nanoparticles are insensitive. Our goal was to explore if overexpression of ferritin, a nontoxic iron-binding protein, can be used for noninvasive magnetic resonance imaging (MRI) of cells transplanted into the infarcted heart. Mouse skeletal myoblasts (C2C12 cells) were engineered to overexpress ferritin. Ferritin overexpression did not interfere with cell viability, proliferation, or differentiation into multinucleated myotubes. Ferritin overexpression caused a 25% decrease in T₂ relaxation time in vitro compared to wild-type cells. Transgenic grafts were detected in vivo 3 weeks after transplantation into infarcted hearts of syngeneic mice as areas of hypointensity caused by iron accumulation in overexpressed ferritin complexes. Graft size evaluation by MRI correlated tighly with histologic measurements (R² = .8). Our studies demonstrated the feasibility of ferritin overexpression in mouse skeletal myoblasts and the successful detection of transgenic cells by MRI in vitro and in vivo after transplantation into the infarcted mouse heart. These experiments lay the groundwork for using the MRI gene reporter ferritin to track stem cells transplanted to the heart.

STEM CELLS have the ability to differentiate into a diverse range of specialized cell types, giving them the potential to dramatically change the treatment of human disease. Heart failure owing to myocardial infarction is a major cause of death worldwide. Cell transplantation using derivatives of adult or pluripotent stem cells is widely investigated as a promising therapeutic approach for heart failure.^1–3 Current methods of studying stem cell engraftment rely heavily on postmortem histologic sampling. There is a lack of suitable imaging technology for serial noninvasive tracking of therapeutic cells after their engraftment. Stem cell labeling with superparamagnetic iron oxide nanoparticles (SPIO) allows labeled cells to be detected by magnetic resonance imaging (MRI) and is commonly used to track stem cell engraftment.^4,5 However, the MRI signal hypointensity caused by those particles does not reflect the actual cell number after several rounds of cell division owing to particle dilution. In addition, MRI signal from iron oxide nanoparticles does not represent only live cells; particles released from dead cells can be phagocytosed by host cells, thereby misleading MRI tracking.^6,7 Both of these reasons preclude monitoring of long-term stem cell engraftment.

Molecular tagging of graft cells by overexpression of nontoxic MRI-detectable probes, such as the iron-binding protein ferritin, is an alternative approach that may solve the cell tracking problem.^8,9 The use of reporter genes for MRI-based cell tracking offers two important advantages over particle-based techniques: (1) gene expression is correlated much more tightly with cell viability than is particle retention and (2) when integrated into the genome, transgene-based reporters are much less susceptible to signal loss through cell division and therefore are uniquely suited for longitudinal monitoring of cell transplants.

The unique structure and properties of ferritin make it valuable as an MRI reporter. Ferritin is the main intracellular iron storage protein and has a globular structure 12.0 nm in diameter, accumulating significant amounts of hydrous ferric oxide iron in its core. ¹⁰ Ferritin can be considered an “endogenous nanoparticle.” The protein shell isolates iron from the cytoplasm, preventing elevation of hydroxyl radical formation.^11,12 It has been shown that ferritin overexpression can be detected by MRI in C6 glioma tumors, ⁸ used for in vivo studies in the mouse brain^9,13 and for imaging of subcutaneous inoculation of undifferentiated mouse embryonic stem cells. ¹⁴

The potential of using ferritin overexpression for noninvasive imaging of stem cells transplanted into an infarcted heart has not been explored. In this study, we aimed to develop a genetically based technique for molecular imaging of ferritin-tagged cells transplanted into infarcted rodent hearts. Our principal hypotheses for this study were as follows: (1) ferritin overexpression is nontoxic for transduced cells, that is, it does not affect cell viability, proliferation, or differentiation; (2) ferritin overexpression in transduced cells is detectable by MRI in vitro and in vivo after transplantation into an infarcted murine heart; and (3) tagging of transplanted cells by ferritin permits accurate quantification of graft size in the heart.

Materials and Methods

Ferritin Expression Vector Design

The murine ferritin heavy-chain complementary DNA (cDNA) with an HA (influenza hemagglutinin) epitope tag (HA-ferritin) was obtained from Dr. Neeman and Dr. Cohen at the Weizmann Institute, Israel. ⁸ BamHI and HindIII double digestion was used for identifying HA-ferritin presence in the pGEM-T vector. HA-ferritin was released from pGEM-T vector backbone using EcoRI restriction sites and was then ligated into the pcDNA3 vector plasmid downstream of the cytomegalovirus (CMV) promoter, thus enabling strong transgene expression and selection of stably transduced cells via neomycin (G418) resistance. DNA sequencing confirmed the fidelity of the construct. Mouse C2C12 skeletal myoblasts were transfected with pcDNA3-HA-ferritin cDNA using FuGENE6 reagent, and cells were cultured on gelatin-coated tissue culture dishes in growth medium (Dulbecco's Modified Eagle's Medium [DMEM], Invitrogen, Carlsbad, CA) supplemented with 20% fetal bovine serum (HyClone, Logan, UT), 2 mM l-glutamine (Invitrogen), and penicillin-streptomycin. Neomycin (G418) was added to the cell culture medium at 1.2 mg/mL to select for stably transduced cells. G418-selected colonies were trypsinized, replated as mixed mass cultures, and maintained in G418 containing growth medium until use. C2C12 subclones were created by dilute plating of cells and isolation of the subclones using cloning rings.

Assessment of C2C12 Proliferation

Cell proliferation was assessed by monitoring the total number of cells plated in six-well plates (5,000 cells per well) during 1 week of growth using a Beckman Coulter Counter. We compared the growth of wild-type and HA-ferritin C2C12 cells both with and without ferric citrate supplementation (1 mM). All calculations were done in duplicate.

Assessment of C2C12 Differentiation

To assess the effect of ferritin overexpression on differentiation of C2C12 cells into multinucleated myotubes, wild-type and transgenic cells were subjected to a myogenic differentiation protocol where growth medium is replaced by DMEM containing 5% horse serum. Cells were maintained in differentiation medium for 7 days and then fixed with ice-cold methanol. Myosin heavy chain was visualized using monoclonal antifast skeletal myosin heavy-chain antibody MY32 (1:400 dilution).

Western Blot Analysis

To assess transgene protein expression, 5 × 10 ⁵ HA-ferritin and wild-type (control) C2C12 cells were lysed, homogenized, and electrophoresed in 12% polyacrylamide gel (30 μg/lane) using the technique described in Golob and colleagues. ¹⁵ Membranes were incubated overnight (4°C) with either anti-HA mouse nonconjugated monoclonal antibody (1:1,000; Covance, Inc., Emeryville, CA) or rabbit nonconjugated monoclonal antiferritin antibody (1:2,000; Abcam Ltd., Cambridge, MA). Identical blots were prepared and incubated with a mouse nonconjugated monoclonal antibody against β-tubulin (1:400; Sigma, St. Louis, MO) as a control for protein loading. Either horseradish peroxidase (HRP)-conjugated goat antirabbit antibody (ferritin detection), sheep antimouse antibody (HA-tag detection), or rat antimouse antibody (β-tubulin detection), all 1:5,000 in Tris-Buffered Saline Tween (TBST), was used as a secondary antibody.

Iron Detection by Prussian Blue Staining

Prussian blue staining was used to confirm iron accumulation in the cytoplasm. C2C12 cells plated in six-well plates were washed twice with phosphate-buffered saline (PBS) and then fixed with 2% paraformaldehyde for 10 minutes. Equal amounts of 20% hydrochloric acid and 10% potassium ferrocyanide solution were mixed and then added to six-well plates for 20 minutes. Nuclear fast red was used as a counterstain. In this assay, the presence of iron was indicated by a bright blue color in a granular cytoplasmic distribution.

Measurements of T2 Relaxation Time by MRI in Cell Samples

To test the capability of MRI to detect ferritin overexpression, T₂ relaxation times were measured in suspensions of wild-type and ferritin-transduced C2C12 cells cultured with and without ferric citrate supplementation (1 mM for 48 hours). For in vitro imaging, wild-type and transgenic C2C12 cells were grown in maxi dishes, washed with PBS to remove excess iron, and then trypsinized and imaged alive in Eppendorf tubes. First, 300 μL of 2% agarose was added to the Eppendorf tubes, and then, after the agarose had cooled off, cells were added in 1 mL of medium (6 × 10 ⁷ per tube). Cells settled down in Eppendorf tubes and formed a loose pellet on top of the agarose. T₂ measurements were obtained from a single slice aligned through the center of the live cell pellets on a 3T Achieva Philips scanner (Seattle, WA) with a custom-built solenoid coil. A multiple spin-echo pulse sequence with 32 equally spaced echoes (10 ms echo spacing) and a repetition time (TR) of 5,000 ms was used. A T₂ map was reconstructed from variable echo time (TE) images by pixel-based fit of a single-exponential signal equation: I = I₀e ^−TE/T2, where I was the signal intensity and proton density (I₀) and T₂ were fitted parameters. T₂ values were measured from the T₂ map in homogeneous regions of interest placed in the center of each sample.

Myocardial Infarction in C3H Mice and C2C12 Injection

All animal procedures described were approved by the University of Washington (Seattle, WA) Institutional Animal Care and Use Committee and performed in accordance with federal guidelines. The C2C12 myoblast line was originally derived from the C3H mouse ¹⁶ ; therefore, we chose this mouse strain as a recipient to minimize immunologic rejection of engrafted cells. Mice were anesthetized by intraperitoneal injection of 2.5% Avertin (Phoenix Pharmaceuticals; 0.02–0.026 mL/g), intubated, and mechanically ventilated with supplemental oxygen and 3 cm H₂O of positive end-expiratory pressure. The heart was exposed via an open thoracotomy and subjected to myocardial injury by permanent ligation of the left anterior descending artery by 8–0 polypropylene suture. After verification that coronary occlusion had occurred (blanching of the tissue distal to the suture), C2C12 cells suspended in 7 μL of serum/antibiotic-free medium were directly injected into the border of the infarcted region of the left ventricle using a 30-gauge needle; 150,000 or 500,000 cells were injected per animal in two injection sites (3.5 μL per site). Six C3H mice received 150,000 ferritin-tagged cells, six animals received 500,000 ferritin-tagged cells, and four mice received wild-type C2C12 cells. The chest was then closed aseptically, and animal recovery from surgery was monitored in a heated chamber.

In Vivo and Ex Vivo MRI of the Murine Heart

Mouse hearts were imaged on a 3T Philips Achieva clinical scanner 3 weeks after the surgery. Mice were anesthetized with 1.5% isoflurane in oxygen (1 L/min) delivered through a nose cone and placed in a solenoid mouse coil (Philips Research Laboratories, Hamburg, Germany) with a built-in heating system maintaining physiologic body temperature. A single-lead electrocardiogram (ECG) was recorded from needle subcutaneous electrodes attached to the animal's extremities and was used to trigger the MRI acquisitions using commercial software (Small Animal Monitoring and Gating System, SA Instrument Inc., Stony Brook, NY).

Each animal was imaged alive and immediately after sacrificing by lethal dose of pentobarbital to confirm the presence of the graft. This approach ensured identification of hypointense MRI signals in transgenic grafts and distinguished true signal from motion and flow artifacts. The in vivo imaging protocol included a two-dimensional ECG-gated T₂*-weighted bright-blood cine turbo-field-echo (TFE) sequence (TR/TE = 15/9.3 ms; flip angle = 15°; slice thickness = 0.8 mm; resolution = 197 × 160 μm) and a black-blood iMSDE (improved motion-sensitized driven equilibrium) prepared TFE sequence (TR/TE = 16/9.8 ms; flip angle = 15°; slice thickness = 0.8 mm, resolution = 197 × 160 μm). The advantage of the iMSDE sequence is improved artifact and blood suppression ¹⁷ and, therefore, more reliable detection of engrafted cells. For graft detection ex vivo, a three-dimensional multiple gradient echo sequence (range 4.9–21.8 ms, echo spacing 4.2 ms) was used with TR = 61.8 ms; flip angle = 10°; slice thickness = 0.5 mm; resolution = 197 × 120 μm; two signal averages.

Histologic Analysis of C2C12 Graft in Mouse Heart

After euthanasia, hearts were fixed in methyl Carnoy solution and processed for histologic analysis. Five micrometer thick tissue sections were cut for histologic staining. Skeletal muscle grafts in mouse hearts were identified using a mouse monoclonal antibody against embryonic skeletal myosin (hybridoma supernatant, 1:100; Developmental Studies Hybridoma Bank, University of Iowa). Sections were blocked with 1.5% normal goat serum in PBS and incubated for 1 hour at room temperature with the biotinylated primary antibody (Animal Research Kit, Dako). Sections were then incubated for 30 minutes at room temperature with HRP-conjugated streptavidin (Dako), developed with 3,3-diaminobenzidine (Sigma), and counterstained with hematoxylin (Sigma). Photographs of heart sections were taken with a QColor 3 Olympus digital camera and a Nikon Eclipse 80i microscope.

Statistical Analysis

Correlation in graft size measurements assessed by MRI was compared to histologic evaluation (embryonic skeletal myosin staining) using Pearson correlation coefficient. Cell proliferation data were analyzed using t-test for unequal variances. Microsoft Excel and SPSS 12.0 (SPSS Inc, Chicago, IL) statistical software were used for analysis.

Results

Ferritin Overexpression in Skeletal Myoblasts

Mouse C2C12 myoblasts were used as a test system to develop the ferritin tagging system. An epitope-tagged version of ferritin heavy chain was stably transduced into C2C12 cells using standard plasmid-based techniques, with transcription driven by the CMV promoter (pcDNA3-HA-ferritin). Stable ferritin transgene overexpression was confirmed by Western blot analysis using mouse monoclonal HA antibody (Figure 1A) and rabbit monoclonal antibody to ferritin (Figure 1B). Jurkat cells and SH-SY5Y cells express high levels of ferritin and served as controls. High levels of ferritin H-chain (21 kDa) were detected in transgenic C2C12 mass culture as well as in a randomly chosen subclone, but not in wild-type cells (see Figure 1B). Prussian blue staining is a sensitive technique for iron detection; it confirmed significant accumulation of iron in the cytoplasm of transduced cells after iron supplementation of the media, whereas no blue cells were observed in iron-supplemented wild-type controls (Figure 1, C and D). Thus, significant overexpression of ferritin was achieved with increased iron storage capacity.

Figure 1.

Confirmation of ferritin overexpression in C2C12 cells. A, Western blot analysis with mouse monoclonal HA antibody indicating ferritin expression in sense clones (1 and 2) of C2C12 cells transduced by pcDNA3-HA-ferritin plasmids but not in antisense clones (3 and 4). B, Western blot analysis using monoclonal rabbit antibody to ferritin: 1 = molecular-weight ladder; 2 = Jurkat cell lysate (ferritin control); 3 = SH-SY5Y cell lysate (ferritin control); 4 = transduced C2C12 cells overexpressing ferritin, mass culture; 5 = transduced C2C12 overexpressing ferritin, subclone; 6 = C2C12 wild type (WT) (negative control). C and D, Prussian blue staining indicates iron accumulation in C2C12 cells transduced by pcDNA3-HA-ferritin plasmids (D), but not in WT cells (C). To facilitate iron loading, cell medium in all wells was supplemented with 1 mM ferric citrate. (original magnification × 60)

Effect of Ferritin Overexpression on C2C12 Viability, Proliferation, and Differentiation

Wild-type C2C12 and cells overexpressing ferritin were grown over 1 month in the standard C2C12 growth medium. Cell viability was evaluated with every cell passage by trypan blue exclusion. No differences in viability were found between wild-type C2C12 and cells overexpressing ferritin. The impact of ferritin overexpression on C2C12 proliferation was assessed by standard growth curves (Figure 2A). The proliferation of the transduced myoblasts was indistinguishable from control myoblasts during 6 days of observation. We noticed that supplementation of cell medium with iron citrate in high doses (1 mM) reduced the expansion of both wild-type C2C12 and transduced cells. However, high iron concentration in the cell medium was more toxic for the wild-type cells than for cells overexpressing ferritin (p < .05). This indicates an advantage for proliferation and/or survival of the ferritin-expressing cells in the presence of high levels of iron. Supplementation of cell medium with low concentrations of ferric citrate (1 μM) did not affect cell growth in either wild-type or transgenic cells (data not shown). To test for possible effects of ferritin overexpression on myoblast differentiation, wild-type and transduced cells were switched from growth medium to differentiation medium and cultured for 7 days. Both wild-type C2C12 and cells overexpressing ferritin differentiated similarly and robustly into multinucleated myotubes, as assessed by immunostaining for fast skeletal myosin heavy chain (Figure 2B). In summary, ferritin overexpression did not interfere with cell viability, proliferation, or differentiation in mouse skeletal myoblasts, and ferritin increased resistance to iron toxicity.

Figure 2.

Similar proliferation rate (A) and differentiation pattern (B) of wild-type (WT) and transgenic C2C12 cells overexpressing ferritin. Cell proliferation was assessed by monitoring the total number of cells plated in six-well plates (5,000 cells per well) during 6 days of growth using a Beckman Coulter Counter. Skeletal myosin heavy-chain immunostaining indicated a comparable differentiation pattern into multinucleated myotubes in WT and transgenic C2C12 cells overexpressing ferritin. Fe supplementation of cell medium by ferric citrate (1 mM).

Changes in Magnetic Relaxation Properties of Transgenic Cells Overexpressing Ferritin

The ferritin-overexpressing cells were readily detectable by MRI in vitro, yielding significant changes in T₂ compared to wild-type cells. T₂ of transgenic cells decreased by ≈25% compared to nonmodified control cells (Figure 3). Supplementation of growth medium by ferric citrate caused additional shortening of T₂ with further amplification of the difference between wild-type and transgenic cells. In summary, these data show that overexpression of ferritin in C2C12 cells can be detected by MRI in vitro.

Figure 3.

In vitro MRI of wild-type (WT) and transgenic C2C12 cells overexpressing ferritin. A, Transverse relaxation (T₂) map of WT cells and C2C12 cells transduced by pcDNA3-HA-ferritin construct with and without iron supplementation (Fe: 1 mM). B, Quantification of changes in T₂ relaxation time of WT C2C12 and cells transduced by pcDNA3-HA-ferritin with and without iron citrate supplementation. Error bars indicate standard deviation of T₂ within region of interest on the parametric map.

MRI Visualization of Ferritin-Tagged Grafts in the Mouse Heart In Vivo and Ex Vivo and Correlation with Histology

To validate the hypothesis that ferritin overexpression is suitable as an MRI reporter for noninvasive imaging of grafted cells, we transplanted wild-type and ferritin-overexpressing myoblasts into the infarcted hearts of syngeneic C3H mice. Both wild-type C2C12 and ferritin-tagged cells successfully grew in mouse hearts and formed skeletal muscle grafts (Figure 4). The presence of transgenic grafts in the infarcted mouse heart was detected by T₂*-weighted MRI as areas of hypointensity caused by accumulation of iron in overexpressed ferritin complexes. Cine MRI techniques did not detect any contractile activity of the area containing skeletal muscle grafts, consistent with previous reports of their lack of electromechanical coupling.^18,19 MRI signal void in the graft area was detected in vivo 3 weeks after transplantation of transgenic cells overexpressing ferritin (see Figure 4A). No signal void areas were detected by MRI in wild-type grafts (see Figure 4C). Embryonic skeletal myosin staining confirmed the presence of skeletal muscle grafts in the left ventricle of mouse heart after transplantation of wild-type C2C12 cells (see Figure 4D), as well as ferritin-overexpressing cells (see Figure 4B). Graft cells were well differentiated and contained sarcomeric structures (see Figure 4E).

Reproducibility of a hypointensity area on bright- and black-blood images provided additional proof of the presence of ferritin-tagged grafts (Figure 5, A and B). The areas of signal reduction were reproducible on both image types and therefore distinguishable from flow artifacts. Our studies showed the persistence of signal void areas in the same short-axis slices of the heart in vivo and ex vivo shortly after sacrificing of the mouse (Figure 5C). An increase in echo time from 4.9 to 21.8 ms on ex vivo images revealed a magnetic susceptibility effect of iron accumulation in the transgenic graft. However, the longer TE compromised the signal to noise ratio (see Figure 5, C1–C5).

The area of signal hypointensity was measured in each MRI at the short-axis plane of the heart. Graft size was assessed as a ratio of graft area to the left ventricle in each slice. The average graft to left ventricle ratio in mouse hearts 3 weeks after transplantation of 150,000 cells was 9.4% (assessed by MRI) and 13.3% (by histology). The average graft to left ventricle ratio after transplantation of 500,000 cells was 22.7% by MRI and 23.1% by histology. Importantly, we found a strong correlation between MRI and histologic measurements of the graft size with a slope of 0.79 and a correlation coefficient R ² = .8 (p < .001; Figure 6). This relationship was not significantly different from the line of unity; the p value for slope is .204 and the p value for intercept is .096, suggesting a lack of bias.

Discussion

The present study demonstrates the feasibility of using the MRI reporter ferritin for noninvasive imaging of cardiac grafts in the mouse heart. The major findings are as follows: (1) stable overexpression of ferritin in C2C12 mouse skeletal myoblasts is feasible; (2) ferritin overexpression does not affect cell viability, proliferation, or differentiation; (3) ferritin-tagged cells are detectable by MRI in vitro, yielding significant changes in signal intensity compared to wild-type cells; (4) the presence of transgenic grafts in the infarcted mouse heart can be detected by MRI as areas of hypointensity; and (5) graft size assessed by MRI correlates well with histologic measurements.

Figure 4.

Detection of the skeletal muscle graft in the infarcted mouse heart. C2C12 cells were not incubated with iron supplement prior to transplantation. A and B, Graft detection in vivo by MRI and ex vivo by embryonic myosin staining 3 weeks after transplantation of 500,000 transgenic C2C12 cells overexpressing ferritin. A stable signal void is revealed in the left ventricle, indicating the presence of transgenic graft. C and D, In vivo MRI and histology of the mouse heart 3 weeks after engraftment of 500,000 wild-type (WT) C2C12 cells. The large skeletal muscle graft is well defined as noncontracting invagination into the left ventricle chamber of the heart. No signal void is detected by T₂* sequence in the area of the WT graft. E, Embryonic skeletal myosin heavy-chain staining of the mouse heart reveals the presence of the skeletal muscle graft. Graft cells are well differentiated and contain sarcomeres.

Stem cell–based cellular cardiomyoplasty is a promising therapy for myocardial infarction. This strategy attempts to enhance cardiac function by repopulating the infarcted region with viable cardiomyocytes and, therefore, holds great promise for restoration of ventricular function. ²⁰ However, low retention and survival of transplanted cells were reported.^21,22 Therefore, development of prosurvival strategies^23,24 and novel imaging techniques to study stem cell engraftmental dynamics is mandatory. Imaging-based cell-tracking methods can potentially evaluate the short-term distribution of infused cells (using iron oxide particles),^4–7 their long-term survival and proliferation (using the gene reporter approach ²⁵ ), and cardiac differentiation (using reporter genes ²⁶ ). The major challenge is to find an optimal agent that allows sufficient contrast from the host tissue and reflects changes in graft size after cell transplantation.

We propose that the natural iron storage protein ferritin will be useful for noninvasive long-term monitoring of transplanted cells into the infarcted heart. It has been shown that overexpression of the ferritin H-chain induces expression of the transferrin receptor and increases iron uptake. ²⁷ Therefore, there are natural mechanisms that shift the iron pool into ferritin-bound storage form; this restores iron homeostasis and prevents iron cytotoxicity. Recent work by Liu and colleagues confirmed that transgenic cells overexpressing ferritin are characterized by an increase in iron content as well as by upregulation of transferrin receptors. ¹⁴

In this study, we overexpressed MRI reporter ferritin in a model cell type (C2C12) and visualized the transgenic cells in vitro and in vivo after transplantation into the infarcted hearts of syngeneic C3H mice. Importantly, ferritin did not interfere with myoblast viability, proliferation, or their differentiation into multinucleated myotubes (see Figure 2). This is important because an increase in free intracellular iron is known to cause production of reactive oxygen species through Harber–Weiss/Fenton reactions, resulting in lipid, protein, and DNA damage.^11,12,28 We noticed that supplementation of cell medium with ferric citrate in high doses (1 mM) inhibited expansion of wild-type and transduced C2C12 cells. However, exogenous iron was slightly more toxic for wild-type cells than for cells overexpressing ferritin (see Figure 2A), suggesting that the transgenic cells may better sequester the excess iron and limit hydroxyl radical formation, resulting in enhanced survival and/or proliferation.

Figure 5.

Reproducibility of detection of transgenic ferritin-overexpressing grafts in the infarcted mouse heart. A, In vivo T₂*-weighted bright-blood turbo gradient echo cine sequence: TR = 15 ms; TE = 9.3 ms; flip angle = 15°; slice thickness = 0.8 mm; acquisition voxel size = 0.25/0.25 mm, reconstruction voxel size = 0.10/0.10 mm; 6 signal averages. B, In vivo two-dimensional iMSDE-prepared (improved motion-sensitized driven equilibrium) black-blood turbo spin-echo pulse sequence: TR = 16 ms; TE = 9.8 ms; flip angle = 15°; slice thickness = 0.8 mm; acquisition voxel size = 0.26/0.26 mm, reconstruction voxel size = 0.10/0.10 mm; 10 signal averages. C, Ex vivo detection of the graft on postmortem images (10 minutes after euthanasia) using multiple gradient echoes. Increase in the TE makes MRIs more T₂*-weighted and thus makes a cellular graft more visible in the mouse heart (arrow). Sequence parameters: three-dimensional gradient echo, TR = 61.8 ms; flip angle = 10°; slice thickness = 0.5 mm; acquisition matrix = 248 × 248; 2 signal averages.

Figure 6.

Correlation between MRI and histologic measurements of graft size. There was good agreement between the two techniques, with the regression line not significantly different from the line of unity. LV = left ventricle.

Ferritin overexpression provided sufficient MRI contrast to make transduced cells detectable in vitro and in vivo in murine hearts. Given that transverse MRI relaxivity of ferritin is much higher than its longitudinal relaxivity, we were able to detect ferritin-tagged cells using T₂*-weighted bright- and black-blood image sequences in a clinical 3 T scanner. It has been shown that transverse relaxivity (1/T₂) linearly increases with an increase in field strength^29,30; therefore, we would expect more effective visualization of ferritin-tagged grafts in higher magnetic field strength. Similarly, visualization of ferritin-tagged grafts should be easier in large grafts in human or large animal hearts.

Our current studies with ferritin-tagged cells are encouraging; however, significant challenges lie ahead. For example, although DNA and protein degradation occur rapidly after cell death, ²¹ there are no data as to how rapidly ferritin complexes undergo degradation after the death of transplanted cells and for how long the MRI signal from iron persists.

It was shown that overexpression of the bacterial iron-binding protein MagA in 293FT cells, ³¹ as well as in N2A cells, ³² resulted in production of magnetic iron oxide nanoparticles, similar to magnetosomes produced by magnetotactic bacteria. Studies by Zurkiya and colleagues demonstrated that cessation of MagA induction resulted in a return of iron content in 293FT cells to control values within 6 days, suggesting activation of iron degradation pathways. ³¹ We expect that iron in overexpressed ferritin complexes would go through similar degradation pathways after cell death; therefore, dead transgenic grafts should lose their paramagnetic properties, rendering the MRI signal from ferritin complexes useful for longitudinal monitoring of viable cells.

When transplanted cells divide, any intracellular label, including iron oxide particles and the ferritin-stored iron would be diluted. In this regard, continuous production of ferritin in daughter cells offers a significant advantage for cell tracking by an MRI-detectable gene reporter over a particle-based cell labeling method. It will be important to determine how quickly ferritin can be produced by daughter cells and how quickly ferritin complexes can bind a sufficient amount of iron from the extracellular environment to be detected by MRI.

In summary, this is the first use of MRI for detection of ferritin gene expression in cardiac grafts. These data indicate that ferritin overexpression can effectively tag stem cells transplanted into the heart. Potential future applications of this technique include studying the dynamics of adult or pluripotent stem cells after transplantation.