Molecular Imaging and Biology

, Volume 13, Issue 6, pp 1204–1214

Impact of Indium-111 Oxine Labelling on Viability of Human Mesenchymal Stem Cells In Vitro, and 3D Cell-Tracking Using SPECT/CT In Vivo

Authors

  • Franz Josef Gildehaus
    • Department of Nuclear MedicineUniversity of Munich (LMU)
  • Florian Haasters
    • Experimental Surgery and Regenerative Medicine, Department of SurgeryUniversity of Munich (LMU)
  • Inga Drosse
    • Experimental Surgery and Regenerative Medicine, Department of SurgeryUniversity of Munich (LMU)
  • Erika Wagner
    • Department of Nuclear MedicineUniversity of Munich (LMU)
  • Christian Zach
    • Department of Nuclear MedicineUniversity of Munich (LMU)
  • Wolf Mutschler
    • Experimental Surgery and Regenerative Medicine, Department of SurgeryUniversity of Munich (LMU)
  • Paul Cumming
    • Department of Nuclear MedicineUniversity of Munich (LMU)
  • Peter Bartenstein
    • Department of Nuclear MedicineUniversity of Munich (LMU)
    • Experimental Surgery and Regenerative Medicine, Department of SurgeryUniversity of Munich (LMU)
Research Article

DOI: 10.1007/s11307-010-0439-1

Cite this article as:
Gildehaus, F.J., Haasters, F., Drosse, I. et al. Mol Imaging Biol (2011) 13: 1204. doi:10.1007/s11307-010-0439-1

Abstract

Purpose

This study investigates the effects of 111In-oxine incorporation on human mesenchymal stem cells’ (hMSC) biology and viability, and the applicability of 111In-oxine for single-photon emission computed tomography/X-ray computed tomography (SPECT/CT) monitoring of hMSC in vivo.

Procedures

HMSC were labelled with 10 Bq/cell. Cellular retention of radioactivity, cell survival, and migration were evaluated over 48 h. Metabolic activity was assessed over 14 days and the hMSC’s stem cell character was evaluated. Serial SPECT/CT was performed after intra-osseous injection to athymic rats over 48 h.

Results

Labelling efficiency was 25%, with 61% of incorporated 111In remaining in the hMSC at 48 h. The radiolabelling was without effect on cell viability, stem cell character, and plasticity, whereas metabolic activity and migration were significantly reduced. Grafted cells could be imaged in situ with SPECT/CT.

Conclusions

111In-oxine labelling moderately impaired hMSC’s functional integrity while preserving their stem cell character. Combined SPECT/CT imaging of 111In-oxine-labelled hMSC opens the possibility for non-invasive sequential monitoring of therapeutic stem cells.

Key words

Mesenchymal stem cells (MSC)111In-oxineRadiolabellingCytotoxicityCell trackingComputed tomography

Introduction

In recent years, cell-based therapeutic strategies have been promoted in the field of regenerative medicine [1, 2]. Preclinical in vivo studies have shown a generally positive effect of infusing undifferentiated stem cells on the regeneration of various tissues, such as bone [3, 4], cartilage [5], and myocardium [6]. Adult human mesenchymal stem cells (hMSC) have properties that may particularly suit them for use in cell-based therapies: relatively easy isolation, rapid proliferation in vitro, and an ability to differentiate along various lineages [79]. Despite optimistic reports on beneficial effects of hMSC-evoked regeneration [1012], their fate upon transplantation is not fully documented. Conventional evaluations in experimental animals entail necropsies, or invasive biopsies, which are not readily available for clinical research.

Ferromagnetically labelled cells can be monitored using magnetic resonance imaging (MRI) [13], but significant limitations are imposed by toxicity of the ferromagnetic particles to their host cells [14]. As an alternative, the fate of radioactively labelled living cells might be followed in vivo through external detection of single-photon emission computed tomography (SPECT). The organometalic complex 111In-oxine has certain properties suited to this task; it is permeable to cell membranes and is retained within living cells by a mechanism of ionic exchange [15] while emitting medium energy gamma photons suitable for external detection. Cell labelling studies have shown 111In-oxine/SPECT to serve for serial monitoring of transplanted cells in vivo [16, 17]. However, as in MRI studies with iron oxide microparticles, it must be considered that 111In-oxine may perturb the labelled cells, potentially altering their viability, metabolic activity, and migration. Of particular concern is the long-term impact of labelling on the cells, given that negative effects mediated by Auger electrons and electromagnetic radiation are likely to be time-dependent and cumulative.

Labelling of hMSC with 111In-oxine has been described in several applications, including treatment of musculoskeletal diseases [16], support of haematopoiesis [18], and gene delivery [19]. Nonetheless, there remains some uncertainty of the longer-term viability and distribution dynamics of implanted hMSC. Therefore, we tested the impact of 111In-oxine on different aspects of the cellular integrity of adult hMSC in vitro, extending over a period of several weeks after labelling. To assess the suitability of this method for imaging in vivo, 111In-oxine-labelled hMSC cells were injected into the marrow of the tibiae of athymic nude rats, and serial SPECT/CT images were obtained at 1, 24, and 48 h after implantation, using a dedicated human system, which had first been tested in a series of phantom studies mimicking the absorption and scatter properties expected for rat studies.

Materials and Methods

Cells and Cell Culture

All experiments were performed using an immortalized, clonally expanded hMSC cell line (SCP-1), over-expressing human telomerase reverse transcriptase via lentiviral gene transfer. The properties of this well-characterized and stable cell line are described in detail elsewhere [20]. All cells were cultured in minimum essential medium alpha l-glutamine (MEMalpha; Invitrogen, Carlsbad, California), supplemented with 10% (v/v) fetal bovine serum (FBS; Sigma, Munich, Germany) and 40 IU/mL penicillin/streptomycin (PAA Laboratories GmbH, Pasching, Austria). Cells culture dishes were maintained in a humidified atmosphere of 95% air with 5% carbon dioxide (CO2) at 37°C. For all experiments, SCP-1 between the 74th and 75th passage were used. Cells were examined using an Axiovert S 100 microscope (Zeiss, Munich, Germany), and photomicrographs were made using a Zeiss black and white digital camera (AxioCam MRm), and processed with the Zeiss Axiovision software.

111In-oxine Labelling

Cells were detached under mild conditions in 0.05% trypsin/0.02% EDTA (Biochrom, Berlin, Germany) for 5 min and centrifuged at 400 g for 5 min. The cell pellet was re-suspended in serum- and glucose-free medium, and re-centrifuged at 400 g for 5 min. For cell counting, the washed pellet was re-suspended in 1 mL of serum- and glucose-free medium; the cell concentration was measured by microscopic examination of 10 μL portions of the suspension.

For labelling experiments, 1 × 104 cells were mixed with 111In-oxine (Covidien, Neustadt, Germany) at doses of 10, 15, 20, and 25 Bq per cell, in 1 mL serum- and glucose-free medium in plastic gamma tubes. Incubations were for 15, 20, and 30 min at 37°C (n = 6 for each radioactivity dose and for each incubation time). Subsequently, the cells were washed twice with phosphate-buffered saline (PBS), and radioactivity concentrations of supernatants and cell pellets were measured in a well counter (Cobra Quantum 5000, Perkin–Elmer, Dreieich, Germany). 111In-oxine-labelling efficiency was calculated by dividing the radioactivity of the supernatant and the pellet (the total radioactivity) by the radioactivity of the cell suspension after washing (final radioactivity).

Since the labelling yield was similar for all radioactivity concentrations, we selected the lower radioactivity concentration (10 Bq/cell) for subsequent experiments, so as to minimize radiotoxicity. Here, suspensions of 5 × 106 cells were mixed with 10 Bq/cell 111In-oxine in 5 mL serum- and glucose-free medium, and incubated for 20 min at 37°C. Next, the cells were washed twice with PBS and the supernatant was collected in tubes for radioactivity measurements, in a dose calibrator (VDC202, Veenstra Instruments, Joure, Netherlands). The cells were then suspended either in growth medium (in vitro experiments) or in PBS (in vivo studies). For the assessment of cellular 111In retention, cells and supernatant of the initial labelling (5 × 106 cells) were separated by centrifugation at 400×g for 5 min (n = 6). Radioactivities in cells and supernatant were measured with a well counter (Cobra Quantum5000). The medium was incubated in the controlled environment as described above, and the percentage of 111In cell retention at 1, 3, 24, and 48 h after labelling was calculated.

Live/Dead Assay

Fluorescence labelling of living cells was obtained using fluorescine diacetate (FDA) and propidium iodide (PI; Sigma-Aldrich Chemie Gmbh, Munich, Germany). Portions containing of 3 × 104 cells collected at 1, 24, and 48 h after 111In-oxine labelling were seeded in triplicate to 12-well cell culture plates, incubated for 1 min with 200 μL FDA/PI staining solution and washed three times with PBS, as described earlier [21]. Image acquisition was performed using an Axiovert S100 (Zeiss, Jena, Germany) microscope equipped with a black and white digital camera (AxioCam MRm; 1,388 × 1,040 pixels), a Zeiss Fluar 10×/0.50 objective a 75 W mercury lamp and appropriate filter sets (AHF, Tübingen, Germany). Three photographs were taken from random positions of each well. Results are shown as percentage of green (live) cells from a sample of 100 cells per well.

WST Assay

A water soluble tetrazolium (WST) assay was performed on both 111In-oxine-labelled cells and non-labelled SCP-1 control cells. For each time point (24 h, 48 h, and 14 days), 3 × 103 cells were seeded in 48-well plates and allowed approximately 5 h to adhere under standard incubation conditions. Then, the cells were incubated with the WST-I reagent (Roche Diagnostics, Mannheim, Germany; 1:10 volumes) and incubated for 4 h. Optical density of the medium was then measured at 450 nm using an ELISA reader. A minimum of two independent experimental runs were performed per incubation time.

Time Lapse Analysis (Scratch Assay)

Cells were plated in six-well cell culture dishes for the experiments lasting 48 h. Using a sterile 200 μL pipette tip, four scratches of approximately 600 μm width were made at right angles in each well. Automated video time lapse microscopy was performed in a heated climate chamber at 37°C with 5% CO2 in a humidified atmosphere. A photograph was taken every 20 min of three scratches in each well, with a cell density of 1 × 104 cells per well. Two independent experiments with 111In-oxine-labelled cells and non-labelled SCP-1 control cells were performed.

Differentiation

111In-oxine-labelled cells were differentiated towards both the osteogenic and the adipogenic lineages, with non-labelled SCP-1 serving as positive controls. As a negative control, both labelled and non-labelled cells were cultured using standard cell culture medium (see above). For osteogenic differentiation, cells were plated in six-well cell culture dishes (NUNC, Wiesbaden, Germany) at a density of 3,000 cells/cm2. Three out of six wells in each plate were incubated with osteogenic differentiation medium, while the other three served as un-induced controls. The differentiation medium consisted of DMEM high glucose medium (PAA Laboratories GmbH, Pasching, Austria) supplemented with FBS, penicillin/streptomycin, dexamethasone, β-glycerol phosphate and l-ascorbic acid bisphosphate (all Sigma-Aldrich Chemie Gmbh, Munich, Germany), as described previously [9]. Osteogenic differentiation was verified using alizarin red staining, according to standard protocols.

Adipogenic differentiation was accomplished as previously described [9] with addition of dexamethasone, indomethacin, insulin, and 3-isobutyl-1-methylxanthine (Sigma) to the standard medium. The maintenance medium contained 0.1 mg/mL insulin in standard medium. Stimulation was initiated after seeding of 6 × 103 cells to each of the 12-wells per culture dish had attained full confluency, and differentiation was assessed by oil red O staining, using a standard protocol.

Immunocytochemical Analysis

Simultaneous multi-colour immunofluorescence staining was performed for the surface markers CD73, CD90 and CD105. Sterile glass slides were seeded with 2 × 104 cells/cm2111In-oxine-labelled or non-labelled SCP-1 cells and cultivated for 72 h. After fixation with 4% paraformaldehyde, cells were blocked 1 h with 4% bovine serum albumin solved in PBS, and then incubated with primary antibodies against CD73 (25 μg/mL; Abgent, USA), CD90 (20 μg/mL; SantaCruz, USA) and CD105 (5 μg/mL; Dianova, Germany), and incubated for 30 min. Thereafter cells were washed three times with PBS for 5 min, and appropriate secondary antibodies were applied; FITC conjugated for CD73, AMCA for CD90 (both Dianova, Hamburg, Germany), and Alexa Fluor 546 for CD105 (Invitrogen, Darmstadt, Germany). Nuclear counterstaining was carried out with DAPI (Invitrogen) applied for 1 min at a concentration of 0.5 μg/mL. Negative controls were performed by omission of the primary antibody.

Animal Preparation

The animal procedure was in accordance with the principles for the care and use of laboratory animals of the Ludwig-Maximilians-University of Munich, and was approved by the “Government Committee of Upper Bavaria” (Regierung Oberbayern). Five female athymic nude rats (200–220 g bodyweight, Harlan, Germany) were anesthetized with an i.p. injection of xylazine (5 mg/kg)/ketamine (100 mg/kg), and maintained with 1.6% isoflurane in oxygen, delivered via a custom-made face mask. Under sterile conditions, rats were positioned in dorsal recumbence, with one hind limb elevated. A longitudinal skin incision was made on the rostral side, proceeding from the distal patella to the tibial tuberosity, and the tibial plateau was exposed by sharp dissection of the patellar tendon. An 18 G needle was inserted into the medullary canal of the tibia between the tibial plateau and the tibial tuberosity. The cavity was flushed with sterile saline solution, and cell suspension of 3–4 × 105111In-oxine-labelled cells (approximately 500–800 kBq) in 20 μL phosphate-buffered saline was injected into the medullary canal using a 30 G needle. The tibia was then sealed with bone wax, and the wound was closed.

SPECT/CT Imaging and Image Analysis

Phantom studies of the double-headed SPECT/CT (Symbia T; Siemens Medical Solutions USA, Inc.) equipped with medium-energy low-purpose collimators, were conducted in order to establish linearity, and accuracy of the attenuation and scatter corrections. The water-filled, cylindrical Plexiglas phantom contained six 50 mL inserts, which were filled with solutions of K2HPO4 (ranging in concentration from 0% to 30%), so as to simulate the bone matrix. The chambers also contained 1.5-mL Eppendorf tubes holding point sources of 200, 300, 600, 800, 1,000, and 2,000 kBq, each with a total volume of approximately 20 μL (Fig. 1). Firstly, an emission recording for counts from the 15% energy windows at both photo-peaks (172 keV and 247 keV) was acquired in a 256 × 256 matrix. A total of 64 projection frames, each lasting 70 s, was acquired over 360°, resulting in a total acquisition time of approximately 40 min. Then the CT scan was obtained with 130 kV, 17 mA and a rotation time of 50 s for attenuation correction of the emission image. Another recording was performed without scatter medium, wherein all Eppendorf tubes were placed directly on the scanning table, thus avoiding interference from scatter within the phantom.
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig1_HTML.jpg
Fig. 1

Fused SPECT/CT images of the cylindrical water-filled phantom used for scatter correction, presented in coronal und transversel slices. Fifty-millilitre inserts were filled with a K2HPO4-solution for mimicking the bone matrix and also contained Eppendorf tubes holding the sources.

The total counts in 3D regions of interest were defined on the three plane views using a tool (Hybrid Viewer, HERMES Nuclear Diagnostics, Stockholm, Sweden) for defining contours with threshold value of 30% of the peak resulting in a volume of interest (VOI) size corresponding to 9 × 9 × 9 mm, equal to, three times the voxel dimension. The total counts of each VOI were plotted against the radioactivity of the point sources. Calibration of the standards was verified relative to the 245.4 keV emission γ-line using a high-resolution HPGe semiconductor detector (Canberra, Meridan, CT) and a reference radiation source calibrated by the DKD (German Calibration Service).

For the in vivo studies, anesthetized rats were positioned in prone position and SPECT/CT whole-body images were acquired immediately after transplantation, and again at 24 and 48 h later. Attenuation and scatter-corrected SPECT data were reconstructed using a Flash-3D algorithm with ten iterations and eight subsets, and smoothing with a 3D spatial Gaussian filter with a full width at half maximum of 6 mm. The CT data were reconstructed with a standard Feldkamp algorithm and B60s kernel and a voxel size of 0.98 × 0.98 × 2 mm, and fusion images were examined in all three planes. 111In-oxine radioactivity was measured in VOIs placed in the tibia.

Statistical Analysis

Statistical analysis was performed using SigmaPlot version 8 (SPSS, Munich, Germany). A value of p < 0.05 according to the Mann–Whitney U test was considered significant.

Results

Labelling Efficiency

Incubation of 1 × 104 hMSC with media containing 10 Bq 111In-oxine/cell for 20 min resulted in mean labelling efficiency of 26 ± 2%, which did not significantly change as radiochemical dose was increased to 25 Bq/cell. With longer incubation there was a substantial decrease of labelling in the 20 and 25 Bq/cell preparations. The percentage of viable SCP-1 cells immediately after radiolabelling was comparable in all groups, but decreased from 92 ± 4% at 10 Bq/cell to 84 ± 11% at 25 Bq/cell, as measured in the Packard gamma counter (Fig. 2). Based on these results, 10 Bq/cell and 20 min incubations were used for subsequent studies. The procedure resulted in an initial specific activity of approximately 2.5 MBq/106 cells. After 1 h in culture, 96 ± 3% of intracellular 111In-oxine retained, which decreased 93 ± 3% at 3 h, 76 ± 5% at 24 h, and 64 ± 4% at 48 h (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig2_HTML.gif
Fig. 2

a Percentage labelling efficiency for 1–2 × 104 SCP-1 hMSC with 10, 15, 20, and 25 Bq 111In-oxine/cell. Incubation for 15, 20, and 30 min yielded mean labelling efficiencies of 20–30% (mean ± SD; n = 6). b Cellular retention of the incorporated 111In at 1, 3, 24, and 48 h after labelling with 111In-oxine. Each point is the mean ± SD of six determinations.

Viability (Live/Dead Assay)

FDA and PI fluorescence microscopy showed survival of 93% of control cells versus 95% of 111In-oxine-labelled cells immediately after labelling, 98% versus 97% at 24 h, and 98% for both groups after 48 h, none of which differences were significant (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig3_HTML.gif
Fig. 3

Lack of effect of 111In-oxine incorporation on the viability of labelled cells at early time points (0, 24, and 48 h). Mean results are presented with 95% confidence intervals for cultures labelled with 111In-oxine at a dose of 10 Bq/cell. There were no statistically significant group differences.

Metabolic Activity and Migration Assay

For analysis of mitochondrial function after 111In-oxine labelling, the metabolic activity of control cells was set to 100%. Relative cell metabolism was significantly decreased to 65 ± 12% after 24 h and to 46 ± 6% after 14 days in culture. Analysis after 48 h revealed a non-significant decrease of metabolic activity to 79 ± 26% (Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig4_HTML.gif
Fig. 4

Effect of 111In-oxine on metabolic activity of SCP-1 cells, assessed by WST assay. Relative to activity in unlabelled control cells (set to 100%, black bar); significant reductions were seen at 24 h and 14 days (P < 0.001).

Endpoints in the scratch assays of motility were enclosure areas of 3 × 105 μm2, approximately 50% of the original area of the scratches. The analysis revealed the time to 50% closure of the artificial wound in the cell culture was significantly (P < 0.001) delayed in the 111In-oxine-labelled cells to a mean ± SD of some 19 ± 4 h in comparison to 25 ± 5 h in the control cultures (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig5_HTML.gif
Fig. 5

111In-oxine labelling reduces the tendency of cells to fill an artificial wound in cell culture monolayer. a An illustration of the progressively filling in a disrupted layer over time 0, 5, 10, 15, 20, and 30 h after creation of the scratch defect. Scale bar equals 200 μm. b Mean time in hours needed for closure of 50% of the original scratch area. Labelled cells took significantly longer to attain this criterion (P < 0.001).

Stability of Stem Cell Character

Positive alizarin red staining in the osteogenic and adipogenic differentiation assays (Fig. 6) showed that untreated and 111In-oxine-labelled stem cells differentiated equally well. Likewise, the oil red O staining showed equivalent adipogenic differentiation. In addition, the expression of the surface markers CD73, CD90, and CD105 was verified immunoytochemically for both treatment groups (Fig. 7).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig6_HTML.gif
Fig. 6

Differentiation assay of 111In-oxine-labelled hMSC (SCP-1). Maintenance of stem cell character is shown by the persistence of capability to differentiate into the adipogenic (ad) and osteogenic (eh) lineages, also seen in unlabelled control cells (a, e), also with corresponding results for undifferentiated control groups with (d, h) and without (b, f) radiolabelling. Bar = 200 μm.

https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig7_HTML.gif
Fig. 7

Immunocytochemical multi-colour staining for cell-surface markers of stem cell character. Representative 111In-oxine-labelled (upper row) and unlabelled (middle row) control stem cells show positive immunofluorescence for CD105 (green), CD90 (red), and CD73 (blue), and triple label overlay images. Results for negative controls (lower row) with omission of the primary antibodies are shown in the lower row. Bar = 50 μm.

Accuracy of SPECT/CT Imaging and In Vivo Imaging

Quantitative SPECT phantom measurements the SPECT data correlated well with calibration standards with (r2 = 0.97) and without (r2 = 0.98) the presence of scatter medium; the maximum deviation was 15% (Fig. 8f). SPECT/CT fusion studies in vivo showed clear emission foci in the tibia of all rats, persisting at least 48 h following application; semi-quantitative VOI analysis showed that the median retention of decay-corrected total counts was 95% (range 96%–94%) at 24 h after implantation and 89% (range 92–87%) at 48 h. There was no evidence of leakage of 111In-labelled cells from the intra-osseous site, and no detectable radioactivity in the lungs, spleen, kidneys, or in the bone marrow distal to the site of injection (Fig. 8).
https://static-content.springer.com/image/art%3A10.1007%2Fs11307-010-0439-1/MediaObjects/11307_2010_439_Fig8_HTML.gif
Fig. 8

a Native CT scan in transversal view, b SPECT/CT overlay in b transversal, c sagittal, and d coronal views of a rat 24 h after implantation of 111In-labelled MSC, acquired with a double-headed Siemens Symbia T apparatus using a medium energy collimator. e Coronal views of SPECT/CT whole-body images from a representative rat scanned first at 1 h and then rescanned at 24 and 48 h after implantation of 111In-oxine-labelled hMSC in the left tibia. f Correlation between SPECT emission counts and activity determined by HPGe-Detector measurements with (triangles) and without scatter medium (circles). Both measurements demonstrate a high agreement between quantitative SPECT imaging and g percentage of the originally injected radiotracer dose detectable in the injection site in the tibia after decay correction. After 48 h more than 89% of the original dose is retained within the tibia (n = 4 animals).

Discussion

Among the most promising methods for monitoring new regenerative medicine strategies is through cell labelling with tracers such as 111In-oxine. The radionuclide decays with a half-life of 2.8 days by a mechanism of electron capture, with a two-peak gamma emission suitable for external detection by SPECT. However, an additional electron ejection process (Auger electron) has considerable potential for toxicity to the host cell. There have been discordant reports on the effects of 111In-oxine on the viability and proliferation capacity of the labelled cells [16, 22, 23], which may furthermore depend on the cell type. In this study, we assessed the vulnerability of hMSC to metabolic perturbation of toxicity from 111In-oxine, employing assays of viability, metabolic activity and migration in vitro, and extending to an initial characterization of cell survival in vivo using SPECT/CT.

The live/dead assay revealed that a labelling dose of 10 Bq/cell and incubation time of 20 min had no adverse effects on the vitality of hMSC cells within the first 48 h. However, the labelling decreased in a time-dependent manner their metabolic and migratory activity in vitro. In particular, impaired mitochondrial activity was evident in the tetrazolium assay in vitro, which progressed during 2 weeks in culture. Earlier studies have mostly assessed the vitality of labelled cells immediately after incubation with the radioactive agent [16, 23], with findings of little initial impairment of cell viability. Others, however, have reported an early negative influence of 111In-oxine labelling on viability and proliferation in stem cells [17, 18]. These discrepancies can partly be explained by dose-dependent effects, as higher 111In-oxine doses are more likely to exert toxic effects, as seen in the present study. In addition, present findings highlight the time dependence of cytotoxic effects, in agreement with earlier studies showing a dramatic decrease in cell viability primarily, between 2 and 7 days after 111In-oxine labelling [22]. Other studies have shown an effect on proliferative capacity of cells at 2 days after labelling, when the radioactivity exceeded a certain threshold [16, 18]. In a study of delayed radiation effects, Jin et al., assessed viability of bone marrow stem cells after incubation with 111In-tropolone over 14 days [23]. There was a clear dose-dependent effect on long-term viability, extending from 100% cell survival with low dose (0.1–0.9 MBq) to complete cell loss (18 MBq). A dose-dependent long-term cytotoxic effect was likewise reported by Gholamrezanezhad et al., with cell death of 80% upon incubation with 7 MBq [22]. Since live/dead assays are imprecise for long-term observation due to artefacts caused by cell passaging or media changes, we performed a WST assay. The reduction of cell metabolism to 46% (Fig. 4), is within the range of long-term cytotoxic effects cited above. Even non-radioactive indium is cytotoxic [24], and moreover, 111In is retained in living cells for up to five half-lives, resulting in substantial decay in situ [25]. Martin, et al., proposed that 111In toxicity is due to high doses of radiation delivered to the cell nucleus from internalized Auger electrons [26].

In addition to different labelling doses, different cell types (both human and non-human) have been used in previous reports, which might be a factor in the observed radiosensitivity. However, present results concur with earlier reports on the emergence of toxicity only after 2 days; the clearly reduced proliferative activity at 2 weeks (five half-lives), suggest the occurrence of cumulative and persistent structural and/or genetic damage. With respect to applications in the field of tissue engineering, our method was suitable for follow-up during only 48 h, due to declining radioactivity. This time interval is also of critical relevance to the biodistribution of MSC in vivo, given the eventual requirement for MSC administration via intravenous injection, as has been proposed for augmented repair of bone defects [27], brain injuries [28], and infarcted myocardium [29]. In a recent study with a myocardial infarction model, up to 70% of the injected MSC had accumulated in the lungs after only 5 min, and was thus unavailable for homing [30]. Given this limitation of intravenous procedures, the intra-osseous delivery of MSC in proximity to the fracture site would present a distinct advantage. We found that the cells remained at their injection site (Fig. 8), without evident dispersal to other tissues. However, cell motility was also impaired; our time lapse analyses showed that labelled cells had reduced capacity to close an “artificial wound” in vitro compared with untreated control cells. In contrast to Nowak et al., who detected a negative effect on both cell vitality and proliferative capacity within 48 h after labelling [17], we found that the fundamental stem cell character was unaffected by integration of 111In-oxine. This was a consistent finding by differentiation assays, and also by immunocytochemical staining for stem cell-specific surface marker profiles (CD73, CD90, and CD105), as characterized elsewhere [31, 32]. Likewise, differentiation assays [10, 33, 34] show that characteristic aspects of hMSC remained unaffected.

Our phantom experiments show that a modern SPECT/CT systems, although designed for human, provides spatial resolution sufficent to allow for the semi-quantitative assessment of focal sources in living animals, even arising from volumes as small as 20 μL [35]. The hybrid system accurately localized signals through 3D co-registration of images with anatomical and biological information [36, 37].

Since a number of studies have described loss of 111In as a function of time in culture [38], it is of some interest that we found attenuation of this loss through application of lower radioactivity burden of the cell preparations, which emphasizes the need for separate assessments of each cell type of potential use in cell therapy. In the previous literature, the amount of radioactivity ranged from 0.2 to 260 Bq/cell [1618, 22, 23], with cultivation of at least 1 × 106 cells, the minimum number required for clinical therapeutic studies with hMSC. We transplanted only some 3 × 105 cells per animal, delivered intra-osseally in a very small volume. By this means we obtained a specific activity of 2.5 Bq/cell, which resulted in good SPECT image quality. This method results in retention of radioactive label (67%) at 48 h in vitro, which was nearly double that seen in an earlier report [18]. Furthermore, the SPECT/CT results revealed nearly complete retention of radioactivity in the implantation site over the observed time period, and no evidence of redistribution to other organs. Any loss of radioactivity is mainly attributable to radioactive decay. Furthermore the detected signal can be attributed to viable cells; in the event of cell death, the radiotracer would be eliminated via the kidneys and the bile duct [18, 39, 40].

Our 111ln-oxine-labelling method proved to be adequate for the monitoring of transplanted cells by serial SPECT/CT imaging for at least 2 days. This would be conducive to a wide range of cell-based research studies. For example, in tissue engineering strategies entailing the application of viable cells in supportive 3D scaffolds, transplantation entails high pressures and pulsatile flow, which might result in loss of viability [41]. Still to be resolved is the ability of transplanted stem cells to migrate to a defect site, such as a bone fracture or a myocardial infarct [42, 43], especially given our finding of impaired motility of the 111ln-oxine-labelled cells in culture. However, the cells were metabolically active, and capable of differentiation along two trajectories. This labelling technique may present advantages over alternate procedures, such as the so-called reporter-gene technique, which requires transfer to the target cell or tissue by a reporter-gene bearing (viral) vector [44, 45], which may present problems in translation to the clinical investigations. The new regulatory framework (Regulation (EC) 1394/2007) applied to advanced therapy medicinal products concerns therapeutic procedures used in regenerative medicine, including all cell-based approaches. The framework did not specify methods for the detection of cells upon transplantation, but the present study helps to illuminate the potential of SPECT/CT for this purpose.

Acknowledgements

This work was supported by a research grant from the Bavarian Research Foundation (Bavarian Research Collaboration for Cell-Based Regeneration of the Musculo-Skeletal System in Old Age, www.forzebra.de). We thank Mr. S. Nowak, Ms. M. Buchwald, Ms. T. Bockhöfer, and Ms. A. Brunegraf for assistance in SPECT/CT image acquisition.

Conflicts of Interest

The authors declare that they have no conflict of interest.

Copyright information

© Academy of Molecular Imaging and Society for Molecular Imaging 2010