|Home | About | Journals | Submit | Contact Us | Français|
To develop a clinically applicable MRI technique for tracking stem cells in matrix-associated stem-cell implants, using the US FDA-approved iron supplement ferumoxytol.
Ferumoxytol-labeling of adipose-derived stem cells (ADSCs) was optimized in vitro. A total of 11 rats with osteochondral defects of both femurs were implanted with ferumoxytol- or ferumoxides-labeled or unlabeled ADSCs, and underwent MRI up to 4 weeks post matrix-associated stem-cell implant. The signal-to-noise ratio of different matrix-associated stem-cell implant was compared with t-tests and correlated with histopathology.
An incubation concentration of 500 µg iron/ml ferumoxytol and 10 µg/ml protamine sulfate led to significant cellular iron uptake, T2 signal effects and unimpaired ADSC viability. In vivo, ferumoxytol-and ferumoxides-labeled ADSCs demonstrated significantly lower signal-to-noise ratio values compared with unlabeled controls (p < 0.01). Histopathology confirmed engraftment of labeled ADSCs, with slow dilution of the iron label over time.
Ferumoxytol can be used for in vivo tracking of stem cells with MRI.
Arthritis is one of the most common causes of disability, affecting approximately 49.9 million individuals in the USA, and resulting in 44 million outpatient visits, 992,100 hospitalizations, US$95 billion direct costs for medical treatment and US$47 billion in indirect costs due to lost earnings [1, 2] . A major challenge in the treatment of cartilage defects due to arthritis is the lack of self-regeneration capacity of the injured cartilage [3, 4] . New therapies based on matrix-associated chondrocyte implants or matrix-associated stem-cell transplants (MASIs) provide a potentially curative therapeutic option [5–10] .
A major barrier for evaluations of the long-term success of engraftment outcomes of different matrix-associated chondrocyte implant and MASI approaches is the current inability to recognize complications of the engraftment process in a timely manner. Complications in the early post-transplant period include loss of stem cells from the transplantation site due to mechanical loss and/or apoptosis [11, 12] . An imaging method that could visualize and monitor the presence of stem cells in MASIs directly, noninvasively and longitudinally in vivo would greatly enhance the ability to develop more successful cartilage regeneration techniques.
Superparamagnetic iron oxide nanoparticles (SPIOs) provide a strong signal effect on magnetic resonance (MR) images and can be internalized into stem cells. MRI is currently the only noninvasive diagnostic test that can provide high-resolution anatomical and functional information of cartilage defects in vivo [5,13–16]. A variety of iron oxide nanoparticle-based stem-cell markers have been previously applied for cell-tracking purposes [15, 17–21] . However, these previously applied cell labels are either not clinically applicable or have been taken off the market [22–24] . The authors propose to utilize the US FDA-approved iron supplement ferumoxytol (Feraheme®, Advanced Magnetics, MA, USA) for stem-cell labeling. This agent is currently used for the treatment of iron deficiency  in patients with anemia. Ferumoxytol provides a strong signal on MR images. Thus, the authors hypothesized that this clinically applicable iron oxide nanoparticle compound could be also used as a stem-cell marker  . To the best of the authors’ knowledge, ferumoxytol is currently the only iron oxide nanoparticle compound that could be directly translated to the clinic and applied for stem-cell MRI in patients via an ‘off-label’ use.
Thus, the goal of this study was to develop an immediately clinically applicable MRI test for in vivo tracking of MASIs based on ferumoxytol-labeling of the transplanted stem cells. By exploiting this novel, immediately clinically applicable cell-tracking technique as a new tool to monitor stem-cell engraftment outcomes noninvasively in vivo, the authors anticipate significantly facilitating the development of successful therapies for cartilage regeneration in patients.
Adipose-derived stem cells (ADSCs) have been recently introduced as a new source for MASIs. Subcutaneous adipose depots, the source for these cells, are abundant and easily accessible, thereby providing a potentially unlimited ADSC reservoir. ADSCs demonstrate similar chondrogenic differentiation outcomes when compared with mesenchymal stem cells (MSCs), which are typically harvested from bone marrow aspirations or biopsies [26–29]. Considering these factors, ADSCs have become an attractive alternative source of stem cells for MASI. ADSCs for this study (generous gift from G Lin and T Lue, University of California, San Francisco, CA, USA) were obtained from male Sprague–Dawley rats as previously described by Ning et al.  . Like MSCs, ADSCs express CD29, CD44, CD71, CD90, CD105/SH2, SH3 and the widely recognized stem-cell marker STRO-1. ADSCs do not express CD31, CD45 and CD106 [30–32] . The hematopoietic cell marker CD34 is present in early, but not late, passages [30,33]. The lack of CD106 on ADSCs is consistent with the origin of these cells from nonhematopoietic tissue  . ADSCs were cultured in DMEM (Invitrogen, CA, USA) supplemented with 10% fetal bovine serum (FBS; Invitrogen), 100 U/ml pencillin and 100 µg/ml streptomycin (Invitrogen) at 37°C in a humidified 5% CO2 atmosphere. At 90% confluency, the ADSCs were trypsinized and either redistributed to new culture flasks or used for experiments.
Ferumoxytol (Feraheme) is an iron supplement that has been FDA approved for intravenous treatment of iron deficiencies in patients with renal failure . Feraheme nanoparticles have a mean hydrodynamic diameter of 20–30 nm and are composed of an iron oxide core (diameter of 6.76 nm ± 0.41) and a hydrophilic carboxydextran coat . The superparamagnetic nanoparticles have a strong signal effect on T1- and T2-weighted MR images, reflected by an r1 relaxivity of 38 mM−1 s−1 and an r2 relaxivity of 83 mM−1 s−1 at 20 MHz (0.47 T) .
As a standard of reference for MR signal effects of ferumoxytol, FDA-approved SPIOs, ferumoxides (Endorem®, Guerbet, Aulnaysous-Bois, France and Feridex I.V. ®, Bayer Healthcare, NJ, USA) were used, which have been extensively applied for cell-labeling and cell-tracking purposes [15,17–21] . However, ferumoxides have been recently taken off the market by the pharmaceutical companies and no longer available. Ferumoxides nanoparticles are composed of an iron oxide core (5–30 nm) and a dextran coat. Ferumoxides nanoparticles have a hydrodynamic diameter of 120–180 nm, and a r1 relaxivity of 40 mM−1 s−1 and an r2 relaxivity of 160 mM−1 s−1 at 0.47 T at 37°C  .
ADSCs were labeled with ferumoxytol and protamine sulfate (American Pharmaceuticals Partners, IL, USA) using a labeling technique adapted from previous techniques described by Arbab et al. [36, 37] . To optimize the ferumoxytol-labeling protocol, samples of 0.5 × 106 ADSCs were incubated with labeling media, consisting of increasing concentrations of 0, 100, 200, 300, 400, 500 and 700 µg iron/ml ferumoxytol and 10 µg/ml protamine sulfate. Serum-free ferumoxytol–protamine complex(ed) labeling media was prepared by incubation of ferumoxytol with protamine sulfate at the aforementioned concentrations in serum-free DMEM for 5 min to allow for complex formation based on a recent published technique from the authors’ laboratory . Cell samples were incubated with this serum-free labeling media for 4 h. Then, 10% FBS was added and cells were further incubated overnight at 37°C and 5% CO2.
Initial results (see below) demonstrated that a concentration of 500 µg iron/ml ferumoxytol and 10 µg/ml protamine sulfate provided maximal ferumoxytol uptake and MR signal effects without impairment in ADSC viability. Cells labeled with this concentration were used for in vitro and in vivo MRI studies.
For comparison, control ADSCs were labeled with ferumoxides via simple incubation using established labeling protocols in the authors’ laboratory [15,17,19]. In total, 0.5 × 106 ADSCs were incubated with 100 µg iron/ml ferumoxides in serum-free media for 4 h, followed by an incubation in FBS-supplemented media overnight at 37°C and 5% CO2. Comparisons with recently published Thu et al.’s technique were also performed as per their published method  .
After completion of incubation procedures, all cell samples were washed three times with phosphate-buffered saline by sedimentation (at 25°C at 1000 rpm for 5 min), resuspended in DMEM and used for in vitro or in vivo investigations.
Triplicate samples of 0.5 × 106 ADSCs, labeled with different concentrations of ferumoxytol as described above, were transferred to test tubes, dissolved in 10 µl agarose (4%; Type VII, Sigma-Aldrich, MO, USA) and placed in a waterbath to avoid artefacts from surrounding air. All cell samples underwent MRI on a 7 T MR scanner (MicroSigna 7.0; General Electric, NY, USA) using a custom-built single-channel transmit/receive partial birdcage (internal diameter: 3 cm) radiofrequency (RF) coil for high-resolution MRI.
Sagittal MR images of the cell samples were obtained with a fast spin echo sequence (repetition time: 3000 ms, echo time: 30 ms) and a multiecho spin echo sequence (repitition time: 4000 ms/echo time: 15, 30, 45 and 60 ms). All MR images were obtained with a feld-of-view of 3.5 × 3.5 cm, a matrix of 256 × 256 pixels, a slice thickness of 0.5 mm and a number of excitations of 16. Operator-defined regions of interest were used to determine the mean signal intensity (SI) of each sample on the multiecho spin echo images. T2 relaxation times and T2 maps were calculated by Cine Tool (GE Global Research, NY, USA) based on the data from the multiecho spin echo images for each concentration of ferumoxytol.
The viability of the cell samples was determined 24-h post labeling and just before implantation in the MASI by the trypan blue exclusion test. Labeled ADSCs as well as nonlabeled controls were exposed to trypan blue and the relative number of nonstained, viable cells to the number of stained, nonviable cells was calculated with the use of an automatic cell counter (Countess® Automated Cell Counter, Invitrogen).
The iron concentration within all test samples was determined with inductively coupled plasma optical emission spectrometry. The samples were mineralized with metal-free hydrochloric acid (Fischer Scientific, ON, Canada) overnight and the obtained solutions were nebulized into an argon plasma. Spectrometric analyses were performed by collaborators at the Environmental Measurement 1: Gas–Solution Analytical Center at Stanford (CA, USA), who were blinded with regard to the content of the samples.
Triplicate samples of ferumoxytol-labeled and unlabeled cells were stained with the Accustain™ Prussian blue kit (Sigma-Aldrich) and post-3,3´-diaminobenzidine (DAB) enhancement with the SIGMA FAST™ DAB with Metal Enhancer kit (Sigma-Aldrich). For further confirmation of intracellular compartmentalization of the applied iron oxide nanoparticles, cells were labeled with fluorescein isothiocyanate-conjugated ferumoxytol and imaged on a Cell-IQ® 2 imager, using Imagen 2.6.0 and Analyser 3.0.1 software (Chip-Man Technologies Ltd, Tempere, Finland). The fluorescent signal of labeled cells and unlabeled controls was recorded and evaluated for significant differences using t-test with a significance level of p < 0.05.
Both labeled ADSCs (with 500 µg/ml ferumoxytol and 10 µg/ml protamine sulfate) and unlabeled controls were detached and centrifuged at 1000 rpm for 5 min. In total, 2.5 × 105 cells were resuspended in 0.5 ml of serum-free chondrogenic differentiation media (consisting of KnockOut™ DMEM/F-12 [Life Technologies, CA, USA], 100 U/ml penicillin, 100 µg/ml streptomycin [Gibco, NY, USA], 1% GlutaMAX™ [Life Technologies], 50 µg/ml l-ascorbic acid 2-phosphate sequimagnesium [Sigma-Aldrich], 100 µg/ml sodium pyruvate [Gibco], 40 µg/ml l-proline [Sigma-Aldrich], 100 nM dexamethasone [Sigma-Aldrich], ITS™ + Premix (BD, MA, USA) final concentration: 5.5 µg/ml transferrin, 10 µg/ml bovine insulin, 5 µg/ml sodium selenite, 4.7 µg/ml linoleic acid and 500 µg/ml bovine serum albumin [BD Bioscience, NJ, USA] and 10 ng/ml TGF-β3 [R&D Systems, MN, USA]), and centrifuged again at 1000 rpm for 5 min. Cells were kept as a pellet and the media was changed every 2 days.
At 2 weeks, the pellets were fixed in formalin (10%; BDH, PA, USA) and the same for xylene (EMD Millipore, Darmstadt, Germany). The pellets were then embedded in paraffin and sliced into 5-µm-thick tissue slices on glass slides. The slides were de-waxed and hematoxylin and eosin (H&E) and Alcian blue staining were performed.
Samples (labeled and unlabeled ADSCs) were embedded in gelatin (10%) and fixed in Karnovsky’s fixative: 2% glutaraldehyde (Cat. No. 16000, Electron Microscopy Sciences, PA, USA) and 4% pFormaldehyde (Cat. No. 15700, Electron Microscopy Sciences) in 0.1 M sodium cacodylate (Cat. No. 12300, Electron Microscopy Sciences) pH 7.4 for 1 h at room temperature (RT). Samples post fixed in 1% osmium tetroxide (Cat. No. 19100, Electron Microscopy Sciences) were cut into 1 mm3 pieces and left for 1 h at RT, washed three times with ultrafiltered water, then en bloc stained at 4°C overnight. Samples were then dehydrated in a series of ethanol (50, 70 and 95%) xwashes for 15 min each at 4°C, then followed by two washes of 100% ethanol at RT and one wash of acetonitrile for 15 min. Samples were infiltrated with EMbed 812 resin (Cat. No.1 4120, Electron Microscopy Sciences) mixed 1:1 with acetonitrile for 2 h followed by two parts EMbed 812 to one part acetonitrile for 2 h. The samples were then placed into EMbed 812 for 2 h then placed into molds and filled with resin. The samples were then placed into a 65°C oven overnight to polymerize.
Samples were trimmed and sectioned between 75- and 90-nm thickness on a Leica Ultracut S (Leica, Wetzlar, Germany), picked up on formvar/carbon-coated slot grids (Cat. No. FCF2010-Cu, Electron Microscopy Sciences) or 100 mesh Cu grids (Cat. No. FCF100-Cu, Electron Microscopy Sciences). Grids were contrast stained for 30 s in 1:1 saturated uranylacetate (~7.7%) to 100% ethanol followed by staining in 0.2% lead citrate for 30 s. Cells were imaged on the JEOL JEM-1400 transmission electron microscope (JEOL Ltd, Tokyo, Japan) at 120 kV and photos were taken using an ORIUS™ SC200 digital camera (Gatan Inc., CA, USA).
The study was approved by the institutional animal care and use committee. In eleven nude athymic female Harlan rats, cartilage defects were created in the distal femur of both knee joints under inhalation anesthesia with 1.5–2% isoflurane in 2 l of oxygen. A medial patellar skin incision was made, the patella was dislocated laterally and a circular osteochondral defect (diameter: 2 mm, depth: 1.5 mm) was created in the distal femoral trochlear groove using a microdrill (Ideal, IL, USA). In these defects, either 7.5 × 105 ferumoxytol-labeled ADSCs (right femur: n = 7), ferumoxides-labeled ADSCs (right femur: n = 4) or unlabeled ADSCs (left femur: n = 11) were implanted, using an agarose scaffold (5 µl, Type VII). The implant location and consistency was confirmed visually. After allowing agarose (4%) to harden (1–2 min), the patella was repositioned and the skin incision was closed by Dermalon™ 6-0 monofilament sutures (Covidien, MA, USA).
Ten rats underwent MRI at the day of ADSC implantation and follow-up MR scans at 2 and 4 weeks after ADSC implantation. One additional rat with implants of ferumoxytol-labeled cells underwent MRI on the day of ADSC implantation and at 2 weeks, and was sacrificed directly after the 2-week scan to obtain histopathologic correlations of imaging data at this time point.
MRI of all knee joints was performed with the same 7 T MR scanner and RF coil used for in vitro studies. Animals were anesthetized with 2% isoflurane and placed supine on a custom-built trough with their knees centered in the RF coil. Sagittal MR images of the rat knee joints were obtained with a fast spin echo sequence (repetition time: 3000 ms; echo time: 30 ms; number of excitations: 16; field of view: 2.5 cm; matrix: 256 × 256 pixels; slice thickness: 0.5 mm).
MR images were analyzed using a DICOM-dedicated image processing software (OsiriX, Pixmeo, Geneva, Switzerland). The SI of the stem-cell transplant and the SI of background noise in front of the knee joint (phase-encoding direction) were measured by operator-defined regions of interest. The signal-to-noise ratio (SNR) of labeled and unlabaled MASI was calculated as: SNR = SITRANSPLANT/SINOISE .
Animals were sacrificed at 2 weeks (n = 1) or at 4 weeks after ADSC implantation (n = 10). The knee joints were explanted, dissected and placed in Cal-Ex II (Fisher Scientific, NJ, USA) for 48–72 h. Cal-Ex II is a mixture of formaldehyde and formic acid that fixes and decalcifies the tissue simultaneously. The specimens were then dissected parasagitally, dehydrated through graded alcohol washes, embedded in paraffin, cut in 5-µm sections and stained with standard H&E to define the morphology of the implant. DAB–Prussian blue (Sigma-Aldrich) stains, performed on deparaffinized sections similar to in vitro studies (see above) were used to detect iron oxide nanoparticles. Anti-CD68 staining specific for ED-1 macrophages (primary antibody: mouse anti-rat CD68, Abcam, MA, USA; secondary antibody: Alexa Fluor® 488 Goat Anti-mouse, Invitrogen, OR, USA) was added on 2- and 4-week time points to detect any macrophage infiltration in and around the defect. FISH was performed using a rat Y-chromosome probe (ID Labs, Ontario, Canada) to detect transplanted male rat ADSCs in female rat knees.
Quantitative data of cell samples incubated with different iron oxide concentrations and evaluated at different time points after labeling were compared using t-tests. SNR data of MASI with iron oxide-labeled and unlabeled ADSCs were tested for significant differences with t-tests. Within each ‘in vivo’ group, changes in SNR data over time were compared with an analysis of variance (ANOVA). t-tests and ANOVA models were computed using the t-test and aov functions in R Foundation for Statistical Computing, Vienna, Austria. Since right and left knees of each rat contained different MASI (right knee: labeled ADSCs, left knee: unlabeled ADSCs), the authors assumed that MR scans of each rat’s knee were independent observations. In order to exclude that data from the same rats were dependent (e.g., different rats metabolized the iron oxide labels at different rates), multilevel models were fit to SNR data using the R package ‘lme4’ version 0.999375-39 with specifications identical to each repeated ANOVA. A variable that identified each rat was added as a random effect to a second model, and the fit of each model was compared. In each case, the model fits were not significantly different. For all analyses, a p-value of less than 0.05 was considered to indicate significant differences between different experimental groups or different time points of observation.
All ferumoxytol-labeled ADSCs demonstrated a strong negative (dark) signal on T2-weighted MR images (Figure 1). Corresponding quantitative T2 relaxation times of ferumoxytol-labeled ADSCs were significantly shortened (equaling a higher MR signal effect) compared with unlabeled controls (p < 0.05; Figure 2). There was no significant difference in MR signal effects of ADSCs labeled with ferumoxytol concentrations below 400 µg/ml and above 500 µg/ml (p > 0.05). T2 times at both 400 and 500 µg/ml concentrations were significantly different from lower concentrations (p = 0.0019, Figure 2). Direct comparisons with Thu et al.’s technique (Figure 3) significantly demonstrated lower T2 relaxation times 24-h post labeling than this study’s technique (Thu et al.’s : 7.09 ± 1.02 ms; this study’s: 11.32 ± 1.14 ms, p < 0.05). However, follow-up studies at 14 days after labeling revealed no significant difference in T2 relaxation values for both techniques (Thu et al.’s: 25.06 ± 0.77 ms; this study’s: 25.4 ± 0.81 ms, p > 0.05).
ADSC labeled with ferumoxytol concentrations of 100–500 µg/ml did not show significant differences in viability compared with unlabeled controls (p > 0.05; Figure 2). However, after incubation with high concentrations (600–700 µg/ml), ADSC viability decreased with borderline significance compared with unlabeled controls (p = 0.053; Figure 2)
All ferumoxytol-labeled ADSCs demonstrated a significantly higher iron content compared with unlabeled controls (p < 0.05). Exposure of ADSCs to increasing ferumoxytol concentrations of 100–500 µg iron/ml led to a steadily increasing iron uptake, followed by a plateau at further increased concentrations (Figure 2).
Histopathology and transmission electron microscopy: DAB–Prussian blue staining and fluorescence microscopy confirmed intracellular iron uptake (Figures 4A, 4B, 4D & 4E). ADSCs labeled with fluorescein isothiocyanate-linked ferumoxytol nanoparticle showed intracellular fluorescence on fluorescence microscopy (Chip-Man Technologies Ltd) with significantly higher fluorescence compared with unlabeled controls (p < 0.05; Figure 4F).Transmission electron microscopy showed iron nanoparticles within secondary lysosomes of the labeled cells (arrows, Figure 4C). Chondrogenic differentiation experiments revealed unimpaired chondrogenesis of labeled cells with positive Alcian blue stains for both labeled and unlabeled cells (Figure 5).
Taken together, a labeling protocol with 500 µg iron/ml ferumoxytol and 10 µg/ml protamine sulfate was considered the best compromise between significant MR signal, maximal iron uptake and preserved viability of labeled ADSCs with uninhibited chondrogenesis. This protocol was used for subsequent in vivo experiments.
Ferumoxytol-labeled MASI demonstrated a marked, negative signal on T2-weighted MR images (Figure 6A). The MR signal characteristics of ferumoxytol-labeled MASIs were similar to ferumoxides-labeled MASIs (Figure 6B). Unlabeled MASIs, on the other hand, demonstrated a relatively high signal on T2-weighted MR scans (Figure 6C). The T2 signal of ferumoxytol-labeled MASIs slowly increased over time while the T2 signal of unlabeled MASIs slightly decreased over time (Figures 6 & 7).
Corresponding SNR data showed significant differences between labeled and unlabeled MASIs directly after the surgery (p < 0.01), and at 2 weeks after MASI (p < 0.05; Figure 7). There was no significant difference in SNR data of ferumoxytol- and ferumoxides-labeled MASIs. At 4 weeks, SNR data of labeled ADSCs (both ferumoxytol and ferumoxides) were not significantly different compared with the unlabeled controls (p > 0.05; Figure 7).
H&E and FISH stains of MASI demonstrated engraftment of ADSC in osteochondral defects (Figures 8A–8F). There was no apparent difference between ferumoxytol- and ferumoxides-labeled and unlabeled ADSC transplants. DAB–Prussian blue showed positive staining for iron oxide nanoparticles at 2 weeks after implantation of labeled cells, which corresponded to MRI data (Figure 9A). Ferumoxytol was not detectable any more at 4 weeks post MASI, presumably due to dilution of the iron oxide label (Figure 9B). Control transplants of unlabeled cells did not show positive staining with DAB–Prussian blue (Figure 9C). No positive CD68 staining was seen in labeled and control transplants suggesting no macrophage infiltration in the transplant (supplementary Figure 1, see online at: www.futuremedicine.com/doi/suppl/10.2217/nnm.12.198). FISH analyses demonstrated multiple areas of red fluorescence that correlated to the Y chromosome present in the nuclei of transplanted cells (counterstained by 4’,6-diamidino-2-phenylindole:blue) in all implants (Figures 8D–8F & 9D–9F), which corresponded to the presence of Y chromosome-containing transplanted cells in female Sprague–Dawley hosts with an XX genotype.
Data showed that ADSCs, labeled with ferumoxytol, could be detected in osteochondral defects of arthritic joints with MRI. Since ferumoxytol is FDA approved as an iron supplement for intravenous treatment of anemias, this nanoparticle compound would be, in principle, readily accessible for in vivo tracking of stem-cell transplants in patients via an ‘off-label’ application  .
This study’s group and others have shown that the superparamagnetic signal effects of ferumoxytol can be utilized to enhance vessels, visceral organs and various pathologies on MR images [20,41,42]. This study’s group was the first to utilize Feraheme as an intravenous contrast agent for MRI of arthritis in an animal model  . Following intravenous injection of ferumoxytol, a significant MR signal enhancement was noted of vascular structures [43–45], inflammatory arthritis  and tumors [41,46] . To the best of the authors’ knowledge, ferumoxytol has not yet been applied for in vivo cell-tracking purposes.
Previous studies by the study’s group and others used the SPIOs ferumoxides [47, 48] or ferucarbotran  for labeling and tracking of human MSCs in arthritic joints. These relatively large SPIOs are spontaneously phagocytosed by stem cells and thus, allow stem-cell labeling via simple incubation [15,17,19,36, 37,47–49] . However, ferumoxides (Endorem and Feridex I.V.) and ferucarbotran (Resovist®, Schering AG, Berlin, Germany) are not produced any more by the pharmaceutical industry. They have also shown to inhibit chondrogenesis in a dose-dependant manner [5,18,50,51]. Instead, ultrasmall SPIOs (USPIOs) are now being produced, which have wider diagnostic applications and improved safety profiles  . In addition, this study’s data showed that ferumoxytol labeling did not affect chondrogenesis (Figure 5). However, USPIOs are not as efficiently phagocytosed by stem cells as SPIOs and thus, cannot be introduced into target cells via simple incubation protocols. Previous studies demonstrated that the USPIO compounds ferumoxtran-10 (Sinerem® or Combidex® [Guerbet, Aulnay-sous-Bois, France]) and ferucarbotran (Resovist) can be introduced into stem cells using transfections agents such as lipofectine  , poly-l-lysine [54, 55] or protamine sulphate  . Since ferumoxytol has similar physicochemical characteristics to these USPIOs, an assisted labeling technique was also applied in order to shuttle ferumoxytol into stem cells. Using a modified protocol originally described by Arbab et al. , ferumoxytol was introduced into stem cells with protamine sulfate as a clinically applicable transfection agent. As shown by the data, the needed concentration for efficient stem-cell labeling with ferumoxytol (500 µg/ml) was much higher compared with ferumoxides (100 µg/ml). This higher concentration of ferumoxytol is needed because of two reasons. First, the smaller size of USPIOs compared with SPIOs leads to lower cellular uptake and lower labeling efficiency of USPIOs compared with SPIOs. Second, the r2 relaxivity (MR signal effect) of USPIOs is markedly lower compared with SPIOs, requiring higher concentrations for MRI detection. A recently published technique by Thu et al. demonstrated that ferumoxytol can be introduced into cells by generating ferumoxytol–heparin–protamine complexes  . The underlying idea is to create clusters of larger nanoparticles, which are taken up more efficiently by the target cells. As confirmed by the data from this study, the ‘Thu technique’ is more efficient than this study’s. However, for this study’s intended application of transplanting stem cells into injured joints, clinicians requested a ‘heparin-free’ labeling technique to avoid potential secondary bleeds or heparin-induced cartilage damage [57, 58] . Stem cells have been shown to exocytose their phagocytosed contrast agent in vivo  . This exocytosis would potentially also affect associated protamine and heparin components. A potential effect of released heparin and heparin–protamine complexes on a local defect milieu has to be investigated in more detail before such a technique could be translated to clinical trials. With this study’s protocol, this uncertain effect of heparin does not need to be considered.
Previous studies demonstrated that stem cells labeled with SPIOs, could be tracked in animal models for several weeks . Ferumoxides-labeled MSCs could be tracked in arthritic joints for 7–12 weeks, while ferumoxytol and ferumoxides-labeled ADSCs in this study could be only tracked for approximately 2 weeks. This difference is apparently due to a more rapid cellular proliferation and higher metabolic activity of our rat ADSCs as opposed to hMSCs  with consecutive more rapid dilution of iron oxide nanoparticles in ADSCs as opposed to MSCs. In accordance with the observations in this study, Peng et al. and Vidal et al. also reported a higher proliferation potential and longer time interval to senescence for ADSCs compared with MSCs [26,61]. Bermen et al. and Küstermann et al. suggested that the metabolization of the iron oxide label and fading signal effect on longitudinal MRI studies can be used as an MR-detectable indicator for the viability of the stem-cell transplant [62,63]. Novel approaches to increase the cellular retention of the ferumoxytol label, in order to provide longer-lasting cell-tracking capabilities with MRI are necessary.
Gadolinium-based MR contrast agents are also FDA approved and can also be introduced into stem cells with the FDA-approved transfection agent protamine sulfate. This study’s group and others have reported that stem cells, labeled with clinically approved gadolinium chelates, can be tracked in arthritic joints [64,65]. Thus, gadolinium labeling would represent a potential alternative to ferumoxytol labeling. Advantages of the ferumoxytol approach are: higher sensitivity of the iron oxide label; visualization on proton-density images, typically applied for cartilage evaluation in a clinical setting and; potential for combining direct USPIO cell labeling with intravenous gadolinium injections.
In conclusion, the presented MRI technique allows for direct, noninvasive in vivo visualization of transplanted stem cells in MASI. Since an FDA-approved cell label and clinically applicable MRI technology was used in this study, the described technique can, in principle, be readily translated to the evaluation of MASI in patients via an ‘off-label’ use of ferumoxytol as a cellular label. Employing this technique in the clinic could help to diagnose transplant failures early after stem-cell implantation, avoid invasive follow-up studies and reassign patients with transplant failures to alternate treatment regimens.
The data from this study showed that the iron supplement ferumoxytol can be used as a cell marker for in vivo stem-cell tracking with MRI. To the best of the authors’ knowledge, ferumoxytol is currently the only FDA-approved iron oxide nanoparticle compound that can be directly translated to the clinic via an ‘off-label’ application for stem-cell tracking. A limited survival of transplanted stem cells due to mechanical loss, apoptosis or immune rejection is a major barrier for successful tissue regeneration outcomes. As novel stem-cell therapies enter clinical trials, it will be increasingly important to monitor placement, retention and integration of the transplanted cells into the target tissue. Ferumoxytol-mediated stem-cell tracking will enable us to overcome the bottleneck of diagnosing transplant failures, help to avoid invasive follow-up studies of lost transplants, and aid in assigning patients with transplant failure to early interventions or alternative treatment options.
The authors thank J Vancil and G Beck for their excellent help with the creation of the figures for this manuscript.
This work was supported by NIH grant R01AR054458 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases to HE Daldrup-Link and a research stipend from the National University of Singapore to H Nejadnik.
Financial & competing interests disclosure
The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Ethical conduct of research
The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.
Papers of special note have been highlighted as:
of considerable interest