Effects of MRI Contrast Agents on the Stem Cell Phenotype

Cell Populations

mMSC from C57Bl/6 mice were obtained from Dr. D Prockop, Tulane University, USA, where they were isolated and shown to differentiate into adipocytes and osteoblasts (33). Isolation and characteristics of rMAPC from Fisher rats has been described elsewhere (5). mESC from 129 mice (R1 line) were received from the Vesalius Research Institute (K.U. Leuven, Belgium). All cell lines were maintained as described previously (48). Prior to use in our studies, mESC were replated feeder free on 0.1% gelatin (Chemicon, Billerica, USA)-coated plates.

Cell Labeling

The following magnetic resonance iron-based contrast agents were used: Resovist® (particle diameter 60 nm; Shering, Munchen, Germany), Sinerem® (particle diameter 20–40 nm), and Endorem® (particle diameter 80–150 nm) (both Guerbet, Roissy, France). Cell labeling was performed by coculture of the cells with (U)SPIOs with and without transfection agents [23 or 388 kDa poly-l-lysine (PLL)] for 4–24 h. The concentration of (U)SPIOs was tested in a range of 20–500 μg total iron per milliliter of culture medium. Final concentrations used were Resovist® (50 μg/ml); Endorem® (348 μg/ml); Sinerem® (500 μg/ml).

Iron Quantification

Cell pellets of labeled and unlabeled cells were collected 24, 48, and 72 h after labeling and mineralized “au bain marie” prior to Induced Coupled Plasma-Mass Spectroscopy (ICp-MS) (Perkin Elmer, MA, USA). Values per tube were calculated as follows: total iron content per pellet = μg total iron per liter x (mass mineralized BM/density mineralized BM) x dilution (according to Guerbet protocol).

Dextran Staining

Cells were fixed with 10% NBF (60% PBS and 40% formaldehyde; Sigma, St. Louis, MO, USA) for 15 min, washed, and incubated for 15 min with PBS (Sigma) + 0.1% Triton (Sigma). Next, cells were rinsed and stained with mouse anti-dextran-FITC (1:1000, Stem Cell Technologies, Vancouver, Canada) overnight. The next day, cells were rinsed and staining was detected using fluorescent microscope (AxioImager, Zeiss, Gottingen, Germany).

Cell Proliferation

Labeled cells were cultured in their specific expansion medium under standard conditions (48). Cells were counted each day using a nucleocounter. Population doubling time (PDT) was calculated as follows: T x ln2/ln(A/A₀), with T = time between two cell counts, A = the number of cells at end, and A₀ = the initial number of cells. We followed the PDT of the labeled cell populations for 7 days.

Cell Differentiation Assays

Differentiation of mMSC to the osteogenic and adipogenic lineage was performed according to standardized methods (13,14,35). Differentiation of rMAPC to the endothelial, neuroectodermal, and hepatic lineage was performed as described previously (48).

Electron Microscopy

Cells were fixed with 3.5% glutaraldehyde (Electron Microscopy Science, Hatfield, USA) for 1 h at 37°C. Cells were postfixed with 1% osmiumtetroxide (Sigma), rinsed, dehydrated, and embedded in araldite (Durcupan, Sigma). Semithin sections (1.5 μm) were cut with a diamond knife and stained lightly with 1% toluidine blue (Panreac, Barcelona, Spain). Semithin sections were re-embedded in an araldite block and detached from the glass slide by repeated freezing (liquid nitrogen) and thawing. The block with semithin sections was cut in ultrathin (0.05 μm) sections with a diamond knife, stained with lead citrate, and examined under a Tecnai spirit electron microscopy (FEI). Photographic images were taken with a Morada camera (Soft Image System, Munster, Germany). For (U)SPIO particle quantifications an average of five ultrathin sections corresponding to different locations were analyzed for every time point and the quantification was carried out by three independent scientists.

Fluorescence-Activated Cell Sorting (FACS)

FACS staining was performed as described previously (48). Unlabeled and (U)SPIO-labeled cells were collected by trypsinization, washed with PBS (Invitrogen) containing 3% FBS, and blocked for 10 min with 5% rat serum (Jackson ImmunoResearch, West Grove, USA) and 10 μl anti-CD16/CD32 (2.4G2) antibody (BD Bioscience). After washing, cells were incubated for 30 min at 4°C in 3% FBS containing conjugated antibodies. Cells were washed once and resuspended in PBS 3% FBS and analyzed by flow cytometry on FACSCanto (BDPharmingen, San Jose, USA). Antibodies used to characterize rMAPC were anti-CD44-FITC (OX-49) and antiCD31-PE (TLD-3A12). Mouse IgG2a,κ (G155–178) and mouse IgG1,κ (MOPC-31C) are respectively used as isotype controls. For mMSC anti-CD44-APC (IM7) and anti-ckit-PE (2B8) were used and rat IgG2b,κ (A95–1) was used as isotype control (all antibodies were from BDPharmingen, San Diego, CA).

Cytogenetics

Evaluation of cell ploidy was performed as described (48). Demecolcine (10 μg/ml in HBSS; Invitrogen) was added to the media until 50% of the cells were detached. Media was removed and the remaining attached cells were collected following rinsing. Cells were centrifuged at 1000 RPM for 5 min, and the pellet transferred to a conical tube with PBS. 0.0075 M KCl was added, pre-warmed to 37°C. After 10 min, cells were centrifuged and the pellet mixed with fixative. Cell suspensions were dropped on the slide and allowed to dry. Slides were stained with 1:5 diluted Wright-Giemsa stain (0.4% w/v in MeOH) in Gurr Buffer (Gibco). The number of chromosomes per cell was enumerated for at least 40 cells under a light microscope (Zeiss).

Photothrombotic Animal Model and Cell Injections

Animal experiments were approved by the bioethics committee of K.U. Leuven (P06098, Leuven, Belgium). Male and female NOD-SCID/gamma c^-/- mice were obtained from the breeding colony in the SPF facility at K.U. Leuven. Male Fisher 344 rats were purchased from Charles River (Wilmington, USA). All animals were housed with access to food and water. For surgery, 3-week-old mice or 8-week-old rats were anesthetized intraperitoneally with ketamine (Ketamine 1000, 75 mg/kg;CEVA Santé Animale, Libourne, France) and mede-tomidin (Domitor, 1 mg/kg; Orion Pharma, Espoo, Finland).

Cortical photothrombosis was induced in eight mice and eight rats. A vertical incision was made between the right orbit and the external auditory canal. The upper part of the temporalis muscle was cauterized so that the muscle could be displaced. Photoillumination with green light (wave length, 540 nm; bandwith, 80 nm) was achieved using a Xenon lamp (model L-4887; Hamamatsu Photonics, Hamamatsu City, Japan) with heat-absorbing and green filters. The irradiation at intensity of 0.68 W/cm² was directed with a 3-mm optic fiber, which was placed on the exposed skin above the middle cerebral artery. Photoillumination was performed for 20 min after intravenous injection of the photosensitizer Rose Bengal (20 mg/kg, Sigma-Aldrich) in a tail vein. Control experiments were performed without injection of the photosensitizer. No photothrombotic injury was detected in control animals.

Injection of GFP⁺ MAPC was performed in eight control mice to establish detectability thresholds and also in eight rats 2 days after photothrombosis. The head of the animals was positioned in a stereotactic head frame (Stoelting, Wood Dale, IL) for stereotactic injection into the striatum. The skull was exposed by a small midline incision and a hole was drilled into the skull in the appropriate location, using bregma as the reference point. Stereotactic coordinates starting from the dura were as follows: lateral +0.20 cm for left injection, −0.20 cm for right injection; anterior-posterior +0.05 cm; depth 3 mm to 2 mm (mice); lateral +0.28 cm for left injection, −0.28 cm for right injection; anterior-posterior 0.11 cm; depth 4 mm (rat). Using a 30-gauge Hamilton syringe (VWR International, Haasrode, Belgium), 10,000 or 50,000 labeled cells were injected in mice or 10,000 or 1,000,000 cells were injected in rat at a rate of 0.25 μl/min into left or right striatum, respectively. After injection, the needle was left in place for an additional 5 min before slowly withdrawn. Anesthesia was reversed with atipamezol (Antisedan, 0.5 mg/kg; Orion Pharma), administered intraperitoneally.

In Vitro MRI Experiments of Labeled Cells

For the assessment of the in vitro visualization of (U)SPIO-labeled cells by MRI, phantoms were built using culture dishes (3.5 cm diameter) filled with agar (Sigma-Aldrich, 0.8% in saline). Drill holes (4 mm diameter of a defined depth) were filled with cell suspensions of known concentration. After solidification of the cell-agar suspensions, drill holes were closed with additional agar. All MR images were acquired using a Bruker Biospec 9.4 Tesla small animal MR scanner (Bruker Biospin, Ettlingen, Germany; horizontal bore, 20 cm) equipped with actively shielded gradients (600 mT m⁻¹). A purpose-built radiofrequency solenoid transmit–receive coil with a length of 6 cm and an inner diameter of 4 cm was used for all phantom experiments. 2D multislice–multiecho (MSME) experiments were acquired for the calculation of T2 maps (TR = 6,000 ms and 10-TE increments of 10 ms, 2562 matrix, 156 × 156 μm in plane resolution, 0.8mm slice thickness). T2^∗ maps were acquired similarly to MSME experiments using a gradient echo pulse sequence and 10-TE increments of 4.5 ms. Three-dimensional (3D), high-resolution T2∗-weighted MR images were acquired using a gradient echo sequence (FLASH, TR = 200 ms, TE = 15 ms, flip angle 30°). The field-of-view was 3.8 × 3.8 × 0.75 cm. The resolution was usually 74 × 74 × 59 μm.

In Vivo MRI Experiments

MR images from animals were usually acquired within 12 h after photothrombosis. Subsequent images were acquired within 12 h after cell engraftment and were repeated for up to 10 days after implantation. MR images were acquired using an actively decoupled linear polarized RF resonator as transmitter (inner diameter 7 cm, Bruker Biospin) with respective mouse and rat brain surface coils (Bruker Biospin) as receiver. After acquisition of 2D multislice localizer images, 2D MSME experiments were acquired for the calculation of T2 maps (TR = 6,000 ms and 10-TE increments of 10 ms, 2562 matrix, 156 × 156 μm in plane resolution, 0.8 mm slice thickness). High resolution 3D FLASH images were acquired for the visualization of labeled cells thereafter, resulting in an isotropic spatial resolution of 100 μm³. Other acquisition parameters were TR = 100 ms, TE = 12 ms, flip angle 30°. Diffusion-weighted MR images (spin echo) were acquired for the first two time points after the photothrombotic injury for assessment of injury. Acquisition parameters were: TR = 1500 ms, TE = 27 ms, B = 1500, in plane isotropic resolution 156 μm², 1 mm slice thickness.

MRI Data Processing

Images were processed using Paravision 4 (Bruker Biospin) and NIH ImageJ. Relaxation rates (r₂) were determined as mean values of homogeneous sections of the cell-loaded areas in the agar phantoms. Values were compared to those of unlabeled cells in the same phantom. Relative quantification was also performed using 3D T2∗-weighted MR images. The relative mean signal intensity of the respective drill holes was determined relative to unlabeled cells (SI = 100%). Data were expressed as mean ± SD.

Immunohistochemistry

Animals were sacrificed at 1.5 or 3 weeks after injection with an intraperitoneal overdose (300 μl/20 g) of pentobarbital (Nembutal; CEVA Santé Animale) and transcardially perfused with 4% (w/v) paraformaldehyde (PFA, Sigma) in PBS. Brains were removed and postfixed overnight in 4% PFA. Serial 50-μm coronal sections were made with a microtome (Vibratome, St. Louis, MO). Hematoxylin and eosin staining was performed for all brains to assess injury. GFP staining was performed for visualization of transplanted cells (1:30, Clonetech, CA, USA).

Statistical Analysis

For in vitro experiments Student's paired two-tailed t-test was used for comparison of two experimental groups. Changes were identified as significant if p < 0.05. When multiple groups were compared, Bonferroni correction was applied. Data are shown as mean ± SD.

Labeling Efficiency of mMSC, rMAPC, and mESC with Resovist®, Sinerem®, and Endorem®

In initial experiments, we tested the optimal time for labeling and whether addition of transfection agents was needed for labeling with Resovist®, Sinerem®, and Endorem®. Greater amounts of (U)SPIOs were taken up by cells incubated for 24 h with either Resovist®, Sinerem®, or Endorem® compared to incubation for 4 h (microscopy after Prussian blue staining). Hence, all subsequent studies were done by incubating cells 24 h. Particle clustering was observed when 23 kDa PLL instead of 388 kDa PLL was used as transfection agent together with Resovist®. Hence, all further experiments were performed using 388 kDa PLL. Comparison of (U)SPIOs uptake with and without the transfection agent (PLL) showed that Resovist® and Sinerem® were taken up (more efficiently) in the presence of PLL compared with uptake experiments without the transfection agent. However, uptake of Endorem® was not affected by the presence or absence of PLL (data not shown). Optimal labeling was defined as being able to detect the lowest cell density (15 cells/μl) for one of the three stem cell populations by MRI of phantoms containing labeled cells. After testing various (U)SPIO concentrations and incubation times, this was achieved using the following conditions: 50 μg/ml Resovist® combined with 0.75 μg/ml PLL (388 kDa), 500 μg/ml Sinerem® combined with 0.75 μg/ml PLL (388 kDa), and 348 μg/ml Endorem® without transfection agent for 24 h (data not shown).

The size, amount, density, and localization of (U)SPIO inclusions in the cells was evaluated by transmission electron microscopy (TEM). As shown in Table 2, the number of inclusions per rMAPC was significantly higher for Sinerem® compared to Endorem® and Resovist®. The average size of inclusions, representing clustered (U)SPIOs, was significantly higher for Resovist® and Sinerem® when compared to Endorem®. The density of inclusions was significantly higher for Endorem® and Resovist® compared with Sinerem®. The quantity of internalized iron per cell was evaluated by ICP-MS as shown for time points day 1, day 2, and day 3 after labeling for 24 h (Fig. 1A–C). When different particles were compared within one cell population, higher concentrations of iron were detected in rMAPC and mMSC labeled with Resovist® (2.51 and 11.58 pg/cell, respectively) at day 1, while in mESC the highest concentration of iron was found following labeling with Endorem® (4.14 pg/cell) at day 1. In general, the most iron was in average detected in mMSC, most likely due to their larger size (surface area measured by TEM: mMSC 363.61 ± 267.76 μm²; rMAPC 157.81 ± 89.93 μm², and mESC 114.54 ± 61.48 μm²).

OPEN IN VIEWER

Figure 1.

Determination of iron content in rMAPC, mESC, and mMSC. (A–C) Iron quantification was performed by ICP-MS. Stem cells were cultured for 24 h with Resovist®, Endorem®, or Sinerem®, washed with PBS twice, and cultured for 1, 2, or 3 days without (U)SPIOs before iron quantification. (A) rMAPC, (B) mESC, (C) mMSC (n = 3); ∗p < 0.016. (D–F) Number of endosomal inclusions over time determined by TEM. mMSC, rMAPC, and mESC were labeled with Resovist®. The amount of inclusions per cell type was measured at time points day 1, day 3, and day 5 after labeling (n = 5).

OPEN IN VIEWER

Table 2.

Statistical Analysis of Number, Size, and Density of Inclusions in rMAPC

	Resovist®	Endorem®	Sinerem®
Number of inclusions	2.93 ± 0.53 (n = 14) ∗	3.36 ± 0.84 (n = 17) ∗	6.35 ± 0.86 (n = 20)
Size of inclusions (μm)	1.74 ± 0.23 (n = 40) ∗	1.20 ± 0.07 (n = 57)	1.46 ± 0.05 (n = 127) ∗
Density of inclusions	2.36 ± 0.21 (n = 40) ∗	2.56 ± 0.16 (n = 57) ∗	1.05 ± 0.02 (n = 127)

Cells were fixed and evaluated under TEM; n = number of samples. Density was measured by visual quantification giving number 0 (as control for unlabeled cells) until 5 (big inclusions). Data are presented as mean ± SE.

∗

p < 0.05.

The quantity of iron per rMAPC decreased significantly at day 3 (fold decrease of 3.9 for Resovist®, 5.3 for Endorem®, and 4.25 for Sinerem®). In mESC there was significant decrease of iron for Resovist® by day 3 (fold decrease 2.1) and for Sinerem® by day 2 (fold decrease 1.4). In contrast, no reduction could be observed in mMSC (see Fig. 1). The number of inclusions over time was also evaluated. As expected, in all three stem cell types, the number of inclusions decreased over time (Fig. 1D–F).

In Vitro Evaluation of Cell Labeling by MRI

Visualization of stem cells by MRI is only successful if the amount of (U)SPIOs taken up is sufficient to alter contrast in MRI. To determine if sufficient particles were present in cells to be visualized by MRI, we made agar phantoms containing different numbers of labeled cells to determine the minimum detectable cell number. To assess the effect of dilution of intracellular (U)SPIOs due to cell proliferation and its influence on cell detectability, phantoms containing cells were scanned 24, 48, and 72 h after labeling. Figure 2 shows 3D T2∗-weighted MRI as well as T2 and T2∗ maps of cell phantoms. Cell densities as low as 5 cells/μl were detectable for cells labeled with Endorem® (mESC, rMAPC, mMSC) or Resovist® (rMAPC, mMSC). Labeling with Sinerem® showed less contrast, resulting in a detectability limit in the order of 75 cells/μl. The MRI contrast strongly depended on the combination of stem cell type and contrast agent. Resovist® was superior over Endorem® for rMAPC and for mMSC labeling. However, labeling of mESC was most efficient using Endorem® compared to Resovist® (Fig. 2A). Moreover, contrast in T2∗-weighted MRI was more pronounced in the larger mMSC for all particles compared to mESC and rMAPC. Absolute quantification of cell numbers was not possible due to large magnetic susceptibility effects of the labeled cells at high densities. Fitting for T2 and T2∗ maps was not possible due to signal quench even for the lowest echo times. Counting of erased pixel clusters in 3D T2∗-weighted MR images of cell suspensions of low density [as described by Kustermann et al. (24)] did not yield a correlation with cell numbers due to cell clustering.

OPEN IN VIEWER

Figure 2.

In vitro visualization of cells by MRI. Agar phantoms were filled with 250, 75, 15, and 5 rMAPC, mMSC and mESC per microliter, following labeling of the cells with Endorem®, Resovist®, or Sinerem®. (A) 3D T2∗-weighted gradient-echo MRI of phantoms loaded with cells labeled for 24 h, followed by 24-h culture in (U)SPIO-free medium. The presence of contrast agent is illustrated by hypointense (dark) contrast. (B) T2 map (left) and T2∗-weighted MRI (right) of Endorem®-labeled mESC. The phantom was loaded with labeled cells at concentration of 75 cells/μl cultured for an additional (1) 1 h, (2) 24 h, (3) 48 h, (4) 72 h, and (5) 120 h in Endorem®-free medium.

When cells were cultured for 48 h in iron-free medium prior to MRI, decreased detectability of labeled cells was observed (Fig. 2B). Consistent with experiments immediately after cell labeling (Fig. 2A), sensitivity of MRI detection strongly depended on the combination of stem cell type and (U)SPIO. In line with the results of iron uptake (Fig. 1), Endorem® was more efficient for mESC labeling whereas Resovist® resulted in better detectability of rMAPC and mMSC. The lower proliferation rates of mMSC compared to rMAPC and mESC resulted in less dilution of the contrast agents with time and similar detectability thresholds after additional 72-h incubation in iron-free medium (data not shown).

Effect of Labeling on Stem Cell Proliferation, Phenotype, and Genetic Integrity

We evaluated whether labeling of mMSC, rMAPC, and mESC with either (U)SPIO affects cell proliferation over a period of 7 days. Results are shown in Figure 3. Labeling of mMSC with Sinerem® resulted in an increased PDT on day 5, compared with unlabeled cells or cells labeled with Endorem® and Resovist®. However by day 7, no significant differences were noted between mMSC labeled with any of the contrast agents compared with unlabeled mMSC (Fig. 3A). Labeling with any of the contrast agents did not affect mESC PDT over the 7-day period (Fig. 3B). By contrast, labeling of rMAPC with either Sinerem® or Endorem® resulted in a lengthening of the PDT on day 5, which was statistically significant compared to cells labeled with Resovist® or unlabeled cells (Fig. 3C). The phenotype of mMSC, rMAPC, and mESC was validated following labeling with the different (U)SPIOs. mESC represent pluripotent stem cells (21), characterized by the expression of key transcription factors that maintain the pluripotent state. Among these is the Pou5f1 transcription factor, also known as Oct4 (32). We determined the transcript levels of Oct4 in mESC 2 days after labeling with the (U)SPIOs. In Resovist®-labeled mESC, the level of Oct4 transcripts was 2.77-fold higher compared to control, although this was not statistically significant (Fig. 4A). We have previously shown that, similar to mESC but in contrast to mMSC, rMAPC also express Oct4, and that presence of Oct4 is associated with the significantly broader differentiation potential of rMAPC (48). Therefore, the expression of Oct4 was also evaluated in rMAPC 2 days after labeling with (U)SPIOs and in unlabeled controls. As for mESC, we could not detect a statistically significant difference in Oct4 transcript levels in rMAPC labeled with any of the (U)SPIOs compared to unlabeled rMAPC (Fig. 4B).

OPEN IN VIEWER

Figure 3.

Population doubling time of rMAPC, mESC, and mMSC labeled with different (U)SPIOs. Population doubling time (PDT = T x ln2/ln(A/A₀), with T= time between two cell counts, A = the number of cells at end, A₀ = the initial number of cells) of cells calculated from day 0 until day 7 after (U)SPIO labeling for the following conditions: control, Resovist®, Sinerem®, Endorem®. (A) mMSC, (B) mESC, and (C) rMAPC (n = 3); ∗p < 0.016.

OPEN IN VIEWER

Figure 4.

Comparison of phenotype of labeled and unlabeled stem cells. mMSC, rMAPC, and mESC were labeled with Endorem®, Resovist®, or Sinerem®. (A, B) Two days after culture in (U)SPIO-free medium, RNA was extracted from rMAPC and mESC and levels of Oct4 transcripts were determined by RT-qPCR. Data are shown as ΔCt (n=3) compared with the housekeeping gene Gadph for (A) mESC and (B) rMAPC. (C, D) Three days after culture in (U)SPIO-free medium, the phenotype of mMSC and rMAPC was evaluated by FACS of cells labeled with antibodies against CD44 and c-kit (mMSC) and CD44 and CD31 (rMAPC). Isotype controls are shown in frame. FACS analysis plot for one of three representative experiments. (C) mMSC, (D) rMAPC.

We and others have shown that mMSC, used in the present study, express among others the cell surface anitgen CD44 at high levels, but not c-kit (33,47). rMAPC are CD44^dim/CD31⁺ (5,48). To ascertain that labeling with (U)SPIOs did not affect the mMSC and rMAPC cell identity, we evaluated cells 3 days after labeling with the three contrast agents by FACS. No obvious changes in cell surface phenotype were identified following labeling (Fig. 4C, D).

We also assessed whether labeling with any of the (U)SPIOs would cause increased aneuploidy of mMSC, rMAPC, or mESC. It should be noted that previous published studies have shown that mMSC become aneuploid very quickly after isolation, including the mMSC population used here (19). Compared to unlabeled control cells, the ploidy of labeled cells was not decreased for up to 5 days after labeling (rMAPC: 72.17 ± 3.50%; mESC: 50.9 ± 8.82%; mMSC: 3.3 ± 0.01%).

Effect of Labeling on Stem Cell Differentiation

Previous reports suggested that labeling of stem cells, such as mMSC, with (U)SPIOs may affect their differentiation ability (23). We therefore tested the effect of (U)SPIO labeling on the differentiation ability of mMSC and rMAPC (Fig. 5).

OPEN IN VIEWER

Figure 5.

Differentiation capacity of mMSC and rMAPC labeled with (U)SPIOs in comparison to unlabeled cells. (A–C) mMSC labeled with Endorem®, Resovist®, or Sinerem® were allowed to differentiate towards osteoblasts (A, B) or adipocytes (C). (A) Osteoblast differentiation was induced using osteogenic medium and progeny of labeled and unlabeled cells evaluated by immunohistochemistry for alkaline phosphatase staining at day 6, and calcium measurement at day 14 (n = 2). (B) Specific transcripts measured by RT-qPCR. ΔCt values of Bsp and Alp are shown. ∗p < 0.05. (C) To induce adipogenic differentiation, labeled and unlabeled cells were cultured with insulin, and progeny evaluated on day 21 by Oil Red O staining. Left: control, right: adipogenic stimulation. Fat vacuoles are formed in the cells. (D–F) rMAPC labeled with Endorem®, Resovist®, or Sinerem® were allowed to differentiate to hepatocyte-like (D), endothelium-like (E), and neuroprogenitor-like (F) cells as described in Materials and Methods. Cells were harvested and on day 21, 9, and 6, respectively, and expression of hepatic, endothelial, and neuroprogenitor transcripts evaluated by RT-qPCR (n = 9). ΔCt values are shown. (G–N) Evaluation of iron inclusions in differentiated rMAPC and mMSC by TEM. rMAPC labeled with Endorem®, Resovist®, or Sinerem® were differentiated towards hepatocyte-like, endothelium-like, and neuroprogenitor-like cells. On day 21, 9, and 6, respectively, cells were evaluated for presence of iron inclusion using TEM. Labeled mMSC were differentiated towards osteoblasts. On day 14 cells were evaluated for presence of iron inclusions using TEM. (G) rMAPC liver Resovist®, (H) rMAPC liver Endorem®, (I) rMAPC endothelium Resovist®, (J) rMAPC endothelium Endorem®, (K) rMAPC endothelium Sinerem®, (L) mMSC osteoblast Endorem®, (M) mMSC osteoblast Resovist®, (N) mMSC osteoblast Sinerem®. Iron particles are seen as black cluster or indicated by arrows.

mMSC differentiate towards osteoblasts and adipocytes (33,35,36). Treatment with dexamethasone, β-glycerophosphate, and l-ascorbic acid, to induce osteogenic differentiation, resulted in a similar increase in alkaline phosphatase (Fig. 5A), and calcium deposition by day 14 when labeled and unlabeled mMSC were compared. Statistically significant increases in transcripts for the osteogenic markers, Alp (26) and Bsp (15), were similar in labeled and unlabeled cells on day 14 of osteogenic differentiation (Fig. 5B). Adipogenic differentiation was induced using insulin and assessed by Oil Red O staining (Fig. 5C). Cells stained with Oil Red O were similar in labeled or unlabeled cells; for Sinerem®-labeled cells a modest decrease in staining is seen based on visual observations. Hence, labeling of mMSC with Resovist®, Sinerem®, or Endorem® did not affect the differentiation ability of mMSC towards osteoblasts and adipocytes. As shown in Figure 5I-N, TEM demonstrated persistent presence of the different (U)SPIOs in the differentiated mMSC progeny over the whole time period of the experiment.

In comparison with mMSC, rMAPC have significantly broader differentiation ability, and can generate progeny of the three germ layers (48). To assess the effect of (U)SPIO labeling on rMAPC differentiation, differentiation to hepatocyte-, endothelium-, and neuroectodermal progenitor-like cells was induced using methods described previously (48). Lineage differentiation was assessed by RT-qPCR for lineage-specific gene transcripts (Fig. 5D–F). Differentiation towards hepatocyte-like cells was induced using a multistep protocol as described previously (48). For labeled or unlabeled cell populations, a similar increase in transcripts for α-fetoprotein (Afp) (9), albumin (Alb) as well as the mature hepatocyte gene Tat (8) was seen by day 21 (Fig. 5D). Differentiation to endothelium of rMAPC labeled with (U)SPIOs and unlabeled controls was induced with VEGF-A. A similar induction of Flk1, Prox1, and Pecam (20) was observed (Fig. 5E). Differentiation of rMAPC towards neuroectodermal progenitor cells was achieved using a protocol based on studies by Conti et al. (10) and described in Ulloa et al. (48). A similar increased expression of transcripts for Sox2 and Pax6 (50) was found in labeled and unlabeled cells (Fig. 5F).

Samples from differentiated rMAPC progeny were fixed and analyzed under TEM. rMAPC-derived hepatocytic and endothelial progeny continued to contain (U)SPIOs in, respectively, 20% and 100% of the cells (Fig. 5G, H).

In Vivo Tracking of Labeled Cells in the Brain

As proof of principle for in vivo cell visualization, rMAPC labeled with Resovist® were stereotactically engrafted into the striatum of NOD-SCID γc^-/- mice; 10,000 and 50,000 rMAPC suspended in 2 μl PBS were injected in the left and right hemisphere, respectively. Animals were monitored repeatedly by MRI for 3 weeks. As shown in Figure 6A, B, 10,000 cells could be visualized in 3D T2∗ MR images. The contrast and the hypointense volume of the implanted cells in MRI remained constant over the observation period (data not shown). Control injections with PBS or unlabeled cells did not change contrast apart from injuries due to the surgical procedure (needle track). Location of cells in the MRI was confirmed by histology (Fig. 6C).

OPEN IN VIEWER

Figure 6.

Determination of in vivo detectability by MR imaging. Different concentrations of rMAPC were labeled with Resovist®, suspended in 2 μl medium and were injected in the striatum of NOD-SCID γc^-/- mice. 3D T2∗-weighted MR images are shown for (A) control animal [coronal and corresponding (dotted lines) axial view with (1) 10,000 labeled MAPC, (2) 10,000 unlabeled MAPC, (3) saline injection, and (4) 2 μl medium + Resovist®] and (B) control animal injected with labeled MAPC [coronal and axial view; (1) left 10,000 cells and (2) right 50,000 cells]. (C, D) GFP staining of cells. (C) Striatum and (D) cells engrafting along the corpus callosum. CC, corpus callosum; Cor, cortex; Str, striatum; ML, midline; SR, stroke region.

As also confirmed by MRI, rMAPC were injected 1–3 mm from a photothrombotic lesion. Cells were found along the corpus callosum towards the infarct region within 24 h. However, no infiltration of rMAPC into the lesion area was seen (Fig. 6D).

Experiments were repeated in 8-week-old Fisher rats with a phototrombotic injury (Fig. 7). Resovist®-labeled rMAPC were injected close to the lesion 24 h after photothrombosis. Combinations of 100,000 labeled (Resovist®) GFP⁺ rMAPC with or without 900,000 unlabeled rMAPC were injected in the brain 2–8 mm from the photothrombotic injury localized in the cortex of male Fisher rats (n = 8). Cell relocation into less dense brain tissue such as the corpus callosum and the stroke region or along the needle tract was visualized by MRI (Fig. 7C). Localization of the cells was confirmed with GFP staining and hematoxylin and eosin staining (Fig. 7D, E). Unlabeled control cells showed a similar distribution pattern suggesting that labeling does not appear to affect the redistribution of rMAPC in vivo (Fig. 7B and data not shown). When mMSC were implanted, no redistribution of the cells along the corpus callosum was found (Fig. 7A).

OPEN IN VIEWER

Figure 7.

In vivo tracking of labeled cells in the rat brain. Cells (100,000) were injected in the brain of rats with photothrombotic lesions (left hemisphere) in the contralateral hemisphere (left). The panels show (left) a diffusion weighted MRI to illustrate the extent of the lesion (acquired immediately after photothrombosis and 24 h before cell engraftment); (middle) 3D T2∗-weighted MRI acquired 24 h after cell engraftment (coronal view); (right) the corresponding axial view. The arrows indicate the injection sites. Images were acquired from (A) an animal with engraftment of 100,000 Resovist®-labeled mMSC (left and right hemisphere; no migration was observed); (B) an animal with engraftment of 100,000 unlabeled rMAPC (left and right); (C) an animal with engraftment of 100,000 Resovist®-labeled rMAPC (left and right; note the migration of cells from the injection site towards the site of photothrombotic injury); (D) GFP staining of tissue slices from animal (C) and (E) hematoxylin and eosin staining of tissue slices from animal (C). (D) and (E) CC, corpus callosum; SR, stroke region; Str, striatum.