MSCs were isolated and expanded from the bilateral femora of adolescent male Sprague-Dawley (SD) rats weighing between 150 and 200 g, as previously described (27). The rats were supplied by the Shantou University Medical College Laboratory Animal Center, and their age was 7–8 weeks. Briefly, the bilateral femora and tibia were harvested and the marrow was flushed out using a syringe filled with Dulbecco’s modified Eagle’s medium (DMEM)/F12 (Gibco, NY, USA) containing 10% fetal bovine serum. The bone marrow was plated into 25-cm² culture flasks and cultured in an atmosphere of 5% CO₂ at 37°C for 48 h. The supernatant containing non-adherent cells was then removed and fresh medium was added. When the cells reached ~80–90% confluence, they were passaged two to three times by repeated trypsinization (0.25% trypsin/0.02% EDTA) (Beyotime Biotechnology Institute, Haimen, China) for 2–3 min and subsequent replating. The MSCs were identified and characterized by the absence of staining for CD45 (type: PE-CD45), a surface marker of hematopoietic stem, and positive staining for CD29 and CD45 (BD Biosciences, Franklin Lakes, NJ, USA). All experimental and animal handling procedures were approved by the Animal Care and Use Committee of Shantou University (Shantou, China).

Gd-DTPA (Magnevist^®; Bayer HealthCare Pharmaceuticals, Montville, NJ, USA) is the standard clinically used MR CA, which has a molecular weight of 938 Da. The effectene transfection reagent (Qiagen, Hercules, CA, USA), which is a non-liposomal lipid transfection reagent, was used to transfect Gd-DTPA into the MSCs. As a result of their negative charge, when mixed with effectene, the Gd-DTPA particles were encapsulated with the cationic lipids, which were then efficiently transferred into cells. For cell labeling, single-cell suspensions of 1×10⁶ MSCs were prepared. Gd-DTPA (30 μl; 0.5 mol/l) and effectene (25 μl) were added to the MSCs in a 25-cm² tissue culture flask containing 5 ml DMEM/F12 medium according to the manufacturer’s instructions and were incubated for 30 min (28). During labeling, the MSCs (1×10⁶ cells) were cultured in tissue culture flasks. Subsequently, the effectene/Gd-DTPA mixture (55 μl) was added directly to the DMEM/F12 culture medium for 24 h incubation under standard cell culture conditions (37°C, 5% CO₂). As a negative control, either 30 μl 0.5 M Gd-DTPA or 25 μl effectene were added to the culture medium of 1×10⁶ cells instead of the 55 μl effectene/Gd-DTPA mixture. Cells were incubated under identical conditions for 4 h in a standard cell culture incubator. A total of 1×10⁶ unlabeled cells were used as a blank control, which were incubated in pure culture medium without the addition of any labeling agents.

Prior to labeling (0 h) and at 24, 48 and 72 h following the initial labeling procedures, the cell viability was determined using trypan blue exclusion assays. Cellular proliferation was determined using an MTT assay, as previously described (29). In brief, 1×10⁴ cells/well were cultured in flat-bottomed 96-well plates. Subsequently, 50% of the MSCs were labeled and the remaining cells remained unlabeled and served as a control. Following labeling, the cells were washed twice using phosphate-buffered saline. Following this, 200 μl complete culture medium was added to each well and the plate was incubated in a CO₂ incubator at 37°C and 5% CO₂. Following 12, 24 and 48 h of incubation, 10 samples of the labeled cells and 10 samples of the unlabeled cells were selected for measurement at each time-point. For measurement, 20 μl MTT (Beyotime Biotechnology Institute, Haimen, China) was added at a final concentration of 5 mg/ml in medium and the cells were incubated for 4 h at 37°C and 5% CO₂. Following incubation, 150 μl dimethylsulfoxide (DMSO; Sigma, St. Louis, MO, USA) was added and the absorbance of the blue formazan product, formed by the enzymatic reduction of MTT by the living cells, was measured at a wavelength of 570 nm (Bio-Tek ELX800, Vermont, USA), with 750 nm as a reference.

Labeling efficiency was measured using spectrophotometry, in which the quantity of Gd-DTPA particles in the cells was determined using a spectrometer (Zeeman Z-8200; Hitachi, Tokyo, Japan), as previously described (30). Electron microscopy (CM-10; Philips Scientifics, Eindhoven, The Netherlands) was performed at 60–80 kV. The cells were evaluated for any structural changes as a result of the labeling procedure and for the presence and localization of intracellular CA particles.

The labeled and unlabeled MSCs were trypsinized, centrifuged and resuspended using 20 μl culture medium in 1.5-ml eppendorf tubes (800r/m, 3 min). The total number of cells in each tube was 5×10⁵. MRI of these tubes was performed at 7.0 T (Agilent 7T MRI; Agilent Technologies, Inc., Santa Clara, CA, USA). The MR pulse sequences included T1-weighted (repetition time TR=400 ms; Echo time TE=9 ms) two-dimensional fast spin echo sequences with a 128×128 matrix, four acquisitions, a slice thickness of 1 mm and a 30×30-mm field of view as well as T2-weighted sequences (TR=2,600 ms; TR=100 ms). In addition, for measurements of T1, a T1 map was obtained with a mixed inversion-recovery spin echo sequence, which was initiated with an inversion-recovery experiment followed by a spin echo experiment (TR/TE=10,000 ms/10 ms) to obtain the T1 map. This sequence was performed with a field of view of 30×30 mm, a matrix of 256×128 pixels and a section thickness of 1.5 mm. The signal intensities of the cell pellets were observed and T1 maps were calculated using image software without further modification, as described previously (31). Each in vitro MSC MRI procedure was repeated eight times.

Adult male SD rats weighing between 250 and 300 g were used to prepare a conventional ischemia model. As described previously (32), ligations of the common carotid artery and external carotid arteries were performed and a suture was placed through the carotid artery into the internal carotid artery.

Following successful construction of the ischemic model, chloral hydrate (Yulonghaizao Company, Qingdao, China) was used to anesthetize the SD rats (3 ml/kg; intraperitoneally). The SD rats were then administered with an injection of 1×10⁶ (10 μl each) Gd-MSCs stereotactically. The injection site was in the cortical area, adjacent to the right middle cerebral artery, at a depth of ~3.5–5.0 mm. The coordinates were 1–l.5 mm posterior to the bregma and 2–2.5 mm lateral to the midline. None of the rats received immunosuppressants.

In order to examine the migration of Gd-MSCs in vivo, 7-Tesla MRI was used to image the cells 1,3, 5 and 7 days after the Gd-MSC injection, using T1-weighted fast spin echo sequences (light/dark cycle 10/14 h, 22–26°C, diet 30–40 g/d, water 60–70 ml/d). Chloral hydrate 1.5ml/kg was used to anesthetize mice intraperitoneally. The scanning parameters were TR/TE 400/9 ms and an average of four. Each image was confirmed by two professionals.

The primary bone marrow stromal cells (BMSCs) were seeded onto petri dishes following spherical suspension in culture medium for 10 h for adherence. After 48 h of adherence, the cells had a stretched appearance with short spindles, triangular centered nuclei and marked refraction, rapidly demonstrating colony of amplification (Fig. 1). The expression of cell surface antigens was detected using flow cytometry. With proceeding incubation time, gradual necrosis was observed in the suspended hematopoietic cells; however, replacing the medium led to the growth of evenly distributed cells with a fusiform fibrotic shape. The expression of cell surface antigens was detected using flow cytometry. The flow cytometeric analysis of the BMSC surface antigens revealed CD29, CD45 and CD90 at 84.69, 0.72 and 90.28%, respectively (Fig. 2).

Gd-DTPA is a small soluble molecule. Following mixing with effectene, numerous Gd-DTPA particles with a negative charge were observed, which were initially condensed together and were subsequently coated with cationic lipids and transferred into cells. The internalized Gd-DTPA particles were mainly aggregated together. Following labeling, the Gd-DTPA particles were clearly visible inside the cytoplasm and accentuated around the cellular apparatus. The high-density particles measured ~50 μm in diameter under the electron microscope (Fig. 3). Using atomic emission spectrophotometry, the effectene-mediated Gd-DTPA labeling efficiency was determined to be 86%.

Evaluation of cell viability by trypan blue exclusion assessment revealed an initial transient reduction in the number of cells. No significant difference was observed in the cell viability between the unlabeled and labeled cells (P>0.05) 24, 48 and 72 h after labeling (Fig. 4). In the MTT-based proliferation assay, no statistically significant decrease was observed in the proliferation of the labeled cells compared with that of the unlabeled MSCs (P>0.05) 12, 24 and 48 h after labeling (Table I).

As shown in Fig. 5, the labeled cells (Fig. 5A) demonstrated markedly higher signal intensity in T1-weighted imaging compared with the unlabeled cells (Fig. 5B) and the cells incubated with either Gd-DTPA (Fig. 5C) or effectene alone (Fig. 5D). This signal feature was consistent with the presence of Gd-DTPA particles inside the cytoplasm. Subsequently, the increased signal intensity of T1-weighted imaging of the labeled cells was evaluated and the T1 values of the labeled cells were further measured. In T1-weighted imaging, the signal intensities of the labeled cells, the cells with Gd-DTPA, the cells with effectene and the unlabeled cells were 1067±34, 635±16, 603±19 and 598±25, respectively. The signal intensity of the labeled cells was increased significantly compared with that of the negative control cells (P<0.001) and the signal intensity of the labeled cells was 1.48 times higher than that of the unlabeled cells. The T1 values of the labeled cells, the cells with Gd-DTPA, the cells with effectene and the unlabeled cells were 886±269, 2,086±58, 2,182±280 and 2,185±200 ms, respectively. The T1 value of the labeled cells was decreased significantly compared with those of the negative controls and the unlabeled cells (P<0.001). The T1 values of the labeled cells were 2.46 times lower than those of the unlabeled cells. In terms of cell proliferation, the T1-weighted image signal intensity of the BMSCs reduced gradually and, at the fifth generation, the signal intensity was similar to that observed in the unlabeled cells (Fig. 6).

Examination of the migration of labeled cells in the in vivo model revealed points of high signal, which became more marked with increasing time. The scope, which was a high signal area in basal ganglia in T1-weighted, also increased gradually with time in the MRI (Fig. 7). The unlabeled control BMSCs lacked these points of increased signal intensity.