In Vivo Tracking of Dual-Labeled Mesenchymal Stem Cells Homing Into the Injured Common Carotid Artery


  • Ai Hong Cao,

    1. Jiangsu Key Laboratory of Molecular and Functional Imaging, Department of Radiology, Zhongda Hospital, Southeast University, Nanjing 210009, China
    Search for more papers by this author
  • Hong Jian Shi,

    1. Jiangsu Key Laboratory of Molecular and Functional Imaging, Department of Radiology, Zhongda Hospital, Southeast University, Nanjing 210009, China
    Search for more papers by this author
  • Yu Zhang,

    1. Department of Biological Sciences and Medical Engineering, Southeast University, Nanjing 210009, China
    Search for more papers by this author
  • Gao Jun Teng

    Corresponding author
    1. Jiangsu Key Laboratory of Molecular and Functional Imaging, Department of Radiology, Zhongda Hospital, Southeast University, Nanjing 210009, China
    • Department of Radiology, Zhongda Hospital, Southeast University, 87 Dingjiaqiao Road, Nanjing 210009, China
    Search for more papers by this author
    • Fax: +86 25 83311083


The aim of this study is to conduct in vivo, noninvasive magnetic resonance imaging of labeled rat bone mesenchymal stem cells (BMSCs) as they home into the site of injured common carotid artery following allograft transplantation. Our study was approved by the Institutional Committee on Animal Research. Purified rat BMSCs were dual labeled with superparamagnetic iron oxide (SPIO) particle and fluorescent DiI dye, and subsequently transplanted into recipient rats injured in the left common carotid arteries. Immediately before and 3 hr, 3, 7 and 12 days after transplantation, the labeled cells were monitored in vivo using a 7T micromagnetic resonance imaging (7T micro-MRI) scanner. The signal-to-noise ratios (SNRs) at the injured sites were corroborated with histological examination using Prussian blue staining and fluorescent imaging. Rat BMSCs were labeled with SPIO and DiI at 100% efficiency. When compared with the baseline level before transplantation, the SNR decreased significantly on Days 3 and 7 after injection in the experimental group (Dunnet t test, P < 0.05), whereas insignificant differences were observed after 3 hr and 12 days (Dunnet t test, P > 0.05). In the control group, no significant differences in SNR were found among different time points (ANOVA, P > 0.05). Histological analyses illustrated that red fluorescence and Prussian blue-positive cells were mainly distributed around the lesion areas of injured common carotid arteries. Rat BMSCs can be efficiently labeled with SPIO and DiI, and the directional homing of labeled cells to the site of injured common carotid arteries after intravascular transplantation could be tracked in vivo with 7T micro-MRI. Anat Rec, 2009. © 2009 Wiley-Liss, Inc.


Stem cells are self-renewing, multipotent, and highly proliferative cells. In recent years, stem cell research has emerged as one of the most promising fields in current biology, and the transplantation of stem cells as treatment for human diseases is a subject with profound significance. When internal organs are injured and exogenous stem cells are introduced into the body, whether these cells can home into the injured site is a critical factor for determining the therapeutic efficacy. However, this research on the homing of transplanted stem cells is largely limited to pathological examination of tissue samples at certain times following transplantation. Such “delayed” examination cannot accurately evaluate the efficacy of stem cell transplantation. Therefore, it is crucial to monitor stem cell behavior in real time in live tissues. One promising method for noninvasive live cell tracking is magnetic resonance (Weissleder,1999; Unger,2003), which permits prolonged imaging duration to monitor dynamic cellular migration, and is especially suitable for in vivo tracking of transplanted cells prelabeled in vitro with MR contrast agents such as superparamagnetic iron oxide (SPIO).

In our previous studies, we have successfully performed 1.5T MR imaging of magnetic labeled cells (Ju et al.,2007; Sun et al.,2008). More recently, 7T micromagnetic resonance imaging (7T micro-MR) specifically designed for small animals was established in our laboratory, enabling the imaging with high contrast and excellent spatiotemporal resolution up to micrometers. In this study, rat bone marrow mesenchymal cells (BMSCs) were isolated and cultured, subsequently dual-labeled with magnetic SPIO nanoparticle and fluorescent DiI dye in vitro. The labeled cells were used for allograft transplantation into recipient animals with acute injuries in the common carotid artery, and 7T micro-MR imaging was performed to study homing of the transplanted stem cells to the injured sites in live animals.



Animal models.

Our study was preapproved by the Institutional Committee on Animal Research. Six-week-old male Sprague-Dawley (SD) rats were used as the donor for cell transplantation, and 7-week-old male SD rats were used as the recipient animals. The SD rats were purchased from the Laboratory Animal Facility of Southeast University Medical School.


The following reagents were used in this study: Dulbecco's modified Eagle's medium (DMEM; Gibco, USA), fetal bovine serum (Hangzhou Sijiqing Biological Engineering Materials, Hangzhou, China), trypsin (Amresco, USA), Prussian blue staining kit (Shanghai Hongqiao Lexiang Medical Reagents, Shanghai, China), and DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindo–carbocyanine perchlorate; Biotium, USA), Trypan blue (Sigma, USA).


The following instruments were used in this study: CO2 incubator (Herabus, Germany), inverted microscope (Zeiss, Germany), fluorescence microscope (Zeiss, Germany), sterile work station (Wujiang Biochemical Purification Equipment Factory, China), cell culture flasks (Corning, USA), and MRI system (Bruker, Biospin MRI PharmaScan 7.0T, 300 MHz, 1H model).


Isolation, culture, and passage of rat BMSCs.

Rat femurs and tibias were obtained under sterile conditions, and bone marrow was flushed out using DMEM medium. Subsequently, single-nucleated cells were enriched by density gradient centrifugation, and inoculated into 50-cm2 culture flasks at 1 × 106 mL−1 in low-glucose DMEM medium with 10% FBS. Cells were incubated in 37°C, 5% CO2 cell culture incubator. Medium change was carried out first after 3 days, and then every 3 days afterward. When attached cells reached 80–90% confluency, cultures were trypsinized and passaged at 1:2. Cell growth was monitored by inverted phase contrast microscope.

SPIO labeling of cells.

Synthesized 8 g/L ferric oxide nanoparticles coupling with poly-L-lysine (Fe2O3-PLL), with core particle size of 15 ± 5 nm and saturation magnetization of 60A*M2/kg, was added to Passage 3 (P3) BMSC cell culture medium to a final concentration of 20 μg/mL for labeling (Ju et al.,2007; Sun et al.,2008). Cells were labeled at 37°C, 5% CO2 incubator for 24 hr, and washed repeatedly with phosphate buffer. A small aliquot of labeled cells were examined by Prussian blue staining on cell culture slides to confirm the presence of intracellular iron particles, and the remaining cells were trypsinized and collected.

DiI labeling of cells.

The SPIO-labeled P3 BMSCs were labeled with 1 μg/mL DiI in low-glucose DMEM without FBS. The mixture was incubated first at 4°C for 15 min, and then at 37°C and 5% CO2 for 2 h. After the medium was removed, cells were washed extensively with phosphate buffer. A small aliquot of cells were used for laser confocal microscopy on cell culture slides to confirm red fluorescent labels in the cytoplasm, and the remaining cells were trypsinized and collected for cell transplantation.

Cell viability.

Trypan blue exclusion assay was carried out to monitor cell proliferation in the labeled and unlabeled groups. Cell suspension (5 μL) was mixed with 5 μL 0.4% Trypan blue solution for 1 min, and 100 cells were counted for each experiment by using a hemocytometer. Cell viability was measured by the Trypan blue exclusion rate (unstained cells/100 cells × 100%).

Rat model for acute common carotid artery injury.

Rats suffering from acute injuries in the common carotid artery were used as the recipient animals for cell transplantation. Prior to injury, splenectomy was first performed on 12 animals anesthetized by intraperitoneal injection of 30 mg/kg sodium pentobarbital. Following 1 week recovery, cervical median incisions were made in these animals. After the left common carotid artery was located and the branch between the external and internal carotid arteries was separated, the distal end of the external carotid artery was ligated to generate occlusion at the proximal end of the left common carotid artery and the distal end of the left internal carotid artery to temporarily block blood flow. Subsequently, the ligature at the distal end of the external carotid artery was lifted slightly, and a pediatric scalp needle was used to puncture the proximal end of the ligature away from the branch point of the common carotid artery. Next, a modified elastic guidewire with 0.014″ diameter was used to pierce the external carotid artery and continued through the arterial branch until it reached the ligature point at the proximal end of the common carotid artery. The guidewire traced back and forth three times to fully peel off the endothelium of the common carotid artery. The guidewire was then withdrawn, and the proximal puncture site of the external carotid artery was ligated. The slip knots at the common and internal carotid arteries were then loosened to allow blood flow. A plastic cuff, measured at 0.5 cm in length and 1 mm in diameter and with a longitudinal incision, was placed outside of the vascular wall, and the neck incision site was sutured. Penicillin was added to 200 K IU/day to combat infection for 3 days. Cell transplantation was carried out immediately and 24 hr after the operation through tail vein injection.

Cell transplantation.

Six animals were transplanted with either BMSCs dual labeled with SPIO and DiI as the experimental group, or unlabeled BMSCs as the control group. Tail vein transplantation was conducted as follows: the animal was secured with special-purpose rodent trap, and the tail was cleaned thoroughly with ethanol to expose the vein. One-milliliter syringe was used to take up 0.2 mL saline, 0.3 mL cell suspension (∼6× 106 cells), and 0.2 mL saline. The tail vein was punctured by the attached 30 g needle, and the cell suspension was slowly injected over 2 min. The puncture point was gently pressed to stop bleeding.


Three animals were randomly chosen at each time point for MRI. The animals were anesthetized with sodium pentobarbital and secured on the examination platform of small animal-specific 7TMicro-MR (Bruker, BioSpin MRI PharmaScan 7.0T, 300 MHz, 1H model). Key parameters were as follows: surface coil (Bruker Biospin) inner diameter, 5.0 cm; field of view (FOV), 3.5 cm × 3.5 cm; excitation times, 2; section thickness, 0.31 mm; imaging sequence: (1) MSME-PD-T2 (modified), 5:1. TR 3,500 ms, TE 20 ms, FA: 180.0°, 512 × 256 (a) matrix, section thickness 0.60 mm, and intersection gap 0.80 mm; (2) MRA: FLASH 3D, 3:1. TR 15 ms, TE 2.5 ms, FA: 20.0°, 256 × 256 × 128 matrix, section thickness 0.31 mm. The signal-to-noise ratio (SNR) at the common carotid artery was measured as follows: the edge of focal lesion at each layer was delineated, and the average signal from blood vessels of that layer was measured and divided by the standard deviation of the background noise to derive the SNR values. The SNR values from 4 to 6 layers were then averaged as the final SNR measurement.

Histological analyses.

One rat was euthanized immediately after MR scan at 3 hr and Days 3 and 7 following transplantation. All the remaining rats were euthanized after the final scan on Day 12. The common carotid arteries were subjected to fast-frozen sectioning for fluorescent microscopy and continuous paraffin sectioning for HE and Prussian blue staining.

Statistical analyses.

SPSS 9.0 software was used for SNR data analysis. The average and standard deviation values were calculated for each time point. The analysis of variance (ANOVA) was performed to compare multiple measurements, and pairwise comparison was conducted using Dunnet t test. P value equal to or smaller than 0.05 was regarded statistically significant.


Morphological Observations of Cultured Cells

BMSCs were isolated, enriched, and expanded using density gradient centrifugation and adherent cell selection. By the first medium change at 72 hr, adherent, spindle-shaped cells were observed. Clonal foci subsequently became visible on Day 4. When the adherent cell density was low, the elongated, spindle-shaped cells grew in random directions, and cells were irregularly aligned. After 12 days, cells became compact; under the microscope, cells were uniformly aligned in bundles and displayed high refraction and robust proliferation (Fig. 1).

Figure 1.

In vitro morphology of cultured rat bone marrow mesenchymal stem cells (BMSCs). After 12 days of in vitro culture, BMSCs grew in a compact, adherent cell layer (P2, ×200).

SPIO Labeling and Prussian Blue Staining

Almost all cells cultured on slides contained blue particles in the cytoplasm after Prussian blue staining. The staining efficiency was 100% (Fig. 2). In contrast, unlabeled cells showed no blue particles.

Figure 2.

Prussian blue staining. Blue particles were observed in the cytoplasm in 100% cells (Prussian blue staining, ×400).

DiI Labeling and Fluorescent Microscopy

All cells following DiI labeling displayed red fluorescence, and the labeling efficiency was 100%. Unlabeled cells showed no red fluorescence (Fig. 3).

Figure 3.

Fluorescent microscopy. Cells were labeled with the fluorescent dye DiI, and red fluorescence was observed in the cytoplasm. The labeling efficiency is 100% (×400).

Effects of Dual Labeling on BMSC Cell Viability

Cell viability was measured by Trypan blue exclusion for both dual-labeled and unlabeled cells. As shown in Table 1, the difference in the exclusion rate between labeled and unlabeled cells was statistically insignificant (P > 0.05) in all time points, indicating that dual labeling did not affect cell growth and proliferation.

Table 1. Trypan blue exclusion rate of the dual-labeled and unlabeled groups in different time
TimeTrypan blue exclusion rate [(x ± s) %]T valueP value
Labeled cellsUnlabeled cells
1 d92.70 ± 3.7193.05 ± 3.820.0810.955
1 w93.20 ± 2.3592.23 ± 2.930.0910.940
2 w92.73 ± 3.6491.85 ± 3.920.1050.928


The SNRs before transplantation and 3 hr, 3, 7, and 12 days after transplantation were measured as follows: for the experimental group, 14.52 ± 2.14, 13.47 ± 1.13, 4.22 ± 2.16, 5.51 ± 2.15, 12.90 ± 1.56, respectively; and for the control group, 14.31 ± 3.12, 15.12 ± 2.23, 14.10 ± 1.79, 14.63 ± 3.47, 14.17 ± 2.82, respectively. In the experimental group, the SNR values at Days 3 and 7 after transplantation showed statistically significant reduction when compared with the SNR measured before transplantation (Dunnet t test; t = 10.03 and 8.01, P < 0.05). These results indicate that SPIO-labeled BMSC cells displayed marked homing into the injured common carotid artery at Days 3 and 7 after transplantation, leading to the reduction of local signal strength. At 3 hr after transplantation, the SNR showed modest reduction, but the difference was not statistically significant (Dunnet t test, t = 1.09, P > 0.05), suggestive of a newly initiated homing of the SPIO-labeled BMSC cells. Likewise, 12 days after transplantation, the signal in the experimental group remained somewhat lower than in the control group, but the difference was not statistically significant (Dunnet t test, t = 1.25, P > 0.05). Possible causes for the reduction in signal strength include the following: (1) labeled cells migrated out of the vascular wall; (2) intracellular iron was diluted after cell division; (3) intracellular iron was secreted out of the cells; (4) cells underwent cell death due to immunological rejection.

In the control group, the SNR values measured after transplantation were not significantly different when compared with the SNR measured before the transplantation with the baseline (ANOVA; F = 0.139, P > 0.05) (Figs. 4 and 5). As expected, MR scan did not detect any changes in signal strength in the intact control common carotid artery before and after transplantation.

Figure 4.

MR imaging of cross sections (PD-T2WI) from the experimental group before transplantation. Robust signal levels were observed on local vascular wall of the injured common carotid artery, as indicated by the arrow.

Figure 5.

MR imaging of cross sections (PD-T2WI) from the experimental group at 3 hr, Days 3, 7, and 12 after transplantation. Statistically significant changes in signal level were observed on local vascular walls at Days 3 and 7, as indicated by the arrow.

Histological Examination

Following Prussian blue staining, positively stained cells mainly distributed around the lesion areas and were observed at all time points (Figs. 6 and 7). On Day 3, the number of stained cells reached maximum, and then declined gradually afterward. Moreover, red fluorescence was observed initially in the injured sites of the common carotid arteries, and gradually migrated toward the median and outer membrane layers over time. In contrast, frozen sections from control animals or the uninjured control arteries did not exhibit any red fluorescence, and no cells were positively stained with Prussian blue.

Figure 6.

Prussian blue staining at Day 3 following transplantation. Blue-stained particles were observed in the lesion area (Prussian blue staining, ×100). The arrows indicate vascular lumen.

Figure 7.

(a) Left panel: Frozen sections of the common carotid artery at Day 7 following transplantation (section thickness: 50 μm). Red fluorescent particles were observed via fluorescent microscopy in the lesion area (×100), as indicated by the arrow. (b) Right panel: after the removal of glycerol and the slide, Prussian blue staining was carried out and blue particles were observed (×100), as indicated by the arrow.


Embryonic or adult stem cells possess totipotent or pluripotent capacity to differentiate into cells of various blastodermic lineages. Extensive animal and clinical studies (Malhi et al.,2002; Balsam et al.,2004; Meyer et al.,2006) have demonstrated that cell transplantation can serve as an innovative therapeutic intervention for cellular restoration and functional repair of injured organs.

SPIO particles exhibit several features ideally suited for MRI: they have small size and strong penetration ability; they can function as MR contrast agent at very low concentrations (nmol); they are safe to use because of their biodegradability and can be metabolized by cells to enter the serum iron pool, where they bind to hemoglobins of red blood cells or participate in other metabolic processes. The magnetic SPIO particles used in this study were prepared by coating the iron oxide nanoparticles with positively charged transfection agent PLL. The resulting complex can enter the cells more readily than larger particles, can be easily absorbed by negatively charged cells, and manifest minimal interference with endogenous cellular processes (Daldrup-Link et al.,2003; Kraitchman et al.,2003; Bos et al.,2004; Tögel et al.,2005). Previously, in our laboratory, we have successfully prepared and applied these magnetic particles in the live tracking of stem cells of distinct lineages using a clinical model of 1.5T MR (Unger,2003; Ju et al.,2006,2007). To ensure safety in this study, we used a low concentration (20 μg/mL) for labeling and still achieved 100% labeling efficiency. Prussian blue staining indicated that Fe2O3-PPL particles were distributed in the cytoplasm of BMSCs and displayed no adverse effects on cell viability and proliferation rate (Unger,2003; Ju et al.,2007). Therefore, this labeling method is safe and effective.

Arbab et al. (2003) argued that following the uptake of dead cells or iron particles by macrophages, iron content would be quickly metabolized or degraded, and 95% of the reduction in local MRI signal may result from loss of labeled cells. However, it is still controversial as to whether the release of iron oxide particles as a result of cell death and lysis may contribute to nonspecific imaging signals. Because iron oxide particles are diluted following cell proliferation, MRI solely dependent on iron oxide is deficient in monitoring stem cell proliferation and differentiation.

Previously, Soriano et al. (1992) discovered that the fluorescent DiI dye can be used to label hepatocytes, providing an excellent morphological means to track the growth and viability of transplanted cells in the recipient tissues. Furthermore, DiI exhibits minimal toxicity toward labeled cells, and they are now widely applied in the labeling of suspension cells. Because of their long lipophilic hydrocarbon chains, DiI molecules can readily stain the cell membrane with great efficiency and stability, generating strong fluorescence in the lipid bilayer. On the other hand, when present in the aqueous phase, they exhibit minimal or weak fluorescence and are essentially devoid of exchange between adjacent cells. In this study, we used SPIO and DiI dual labeling to allow live tracking of fluorescent cells in the recipient tissue in conjunction with MR imaging by 7TMicro-MR, providing accurate description of dynamic changes in morphology and migration of transplanted stem cells.

Current studies on the homing of stem cells mainly involve immunofluorescent and immunohistochemical analyses of tissue sections from animals euthanized at discrete times following transplantation. Although preliminary animal experiments and clinical trials have demonstrated the utility of stem cell transplantation in tissue repair, the process remains poorly defined, primarily, due to lack of understanding of the homing mechanism. Moreover, the inability to dynamically monitor the viability of transplanted stem cells in real-time precludes identification of the optimal time window, route, cell type, and disease for stem cell transplantation, severely hampering clinical applications of stem cells. It is thus important to devise a method to allow noninvasive, real-time tracking of transplanted cells in live tissues (Hill et al.,2003; Bengel et al.,2005).

Recently, a number of reports described labeling and tracking of cells through the use of SPIO nanoparticles or fluorescent dyes (Soriano et al.,1992; Hill et al.,2003; Kraitchman et al.,2003; Werner et al.,2003; Bos et al.,2004; Kong et al.,2004a,b; Corti et al.,2005; Sun et al.,2005,2008; Ju et al.,2007). Moreover, the small animal in vivo imaging systems permit sensitive and quantitative spatial monitoring of markers in whole animals. However, if SPIO labeling alone is used for live tracking, a reduction in signal as revealed by MRI of local injured sites may result from either transplanted stem cells or the presence of hemosiderin during hemorrhage. Prussian blue staining cannot distinguish these possibilities because it detects iron particles regardless of the source. On the other hand, fluorescent labeling can verify the presence of stem cells in the injured site; however, this method is not applicable to live tracking in animals or humans as specimen preparation requires euthanization. In this study, we combined both SPIO and fluorescent dye labeling to monitor the homing of transplanted BMSC cells. MRI live tracking allowed the detection of local signal changes as an indication of BMSC homing; subsequently, the putative homing site in the same specimen was immediately subjected to Prussian blue staining and fluorescence staining via fast-frozen sectioning to verify the homing of BMSC cells. Therefore, fluorescent labeling serves to substantiate the findings from MRI live tracking and to eliminate the complicating factor of hemorrhage. To our knowledge, such dual labeling has not been reported in previous literature.

Molecular imaging techniques have recently been shown to directly or indirectly monitor and record spatiotemporal patterns of molecular and cellular processes, allowing versatile applications in biology, diagnosis, and therapy. In particular, MR-based molecular imaging displays excellent spatial resolution (10–100 μm), soft tissue resolution (Corti et al.,2005), and tissue penetration depth. Thus, MR imaging has become the method of choice to evaluate viability, distribution, and function of transplanted cells (Wang et al.,2006). In this study, we used high-field, small animal-specific MR to perform live tracking of transplanted cells and showed precise signal variation on vascular walls. Specifically, we showed a gradual decrease in the signal strength on the vascular wall of the lesion site on Days 3 and 7; on Day 12, the signal, although still somewhat lower, was not significantly different relative to the baseline.

Extensive studies have demonstrated that in response to organ injury, bone marrow rapidly mobilizes stem cells of multiple lineages to migrate to the damage site, where stem cells are induced by local microenvironment to differentiate into cell types required for tissue repair, thereby promoting recovery of injured tissue. This phenomenon is called “site-specific differentiation,” but the underlying mechanisms are largely unknown. In this study, we showed that the BMSCs introduced via extrinsic pathways into the recipient animals displayed clear chemotaxis, as they appeared to aggregate on the injured sites. Two hypotheses have been proposed with respect to the homing of stem cells. First, necrotic cell death following injury triggers the release of a cascade of signals, which in turn results in the migration of bone marrow stem cells into peripheral blood. In the mean time, the damaged tissue expresses specific receptors or ligands to attract stem cells to the lesion area. Second, stem cells dynamically circulate among different tissues under normal conditions and exit blood circulation in response to tissue damage. Both hypotheses consider the bone marrow pool as the source of circulating stem cells. Researches to date have implicated several signals that may be responsible for the mobilization and homing of stem cells during injury: stromal cell-derived factor-1 (SDF-1) and its receptor CXCR4, stem cell factor (SCF) and its receptor c-kit, clone-stimulating factor (CSF), vascular endothelial growth factor (VEGF), and integrin. Based on our results, we hypothesize that necrotic cells at the injured vascular site release certain signals, eventually leading to local aggregation of exogenous stem cells (Figs. 4–6). Our results provide animal-based experimental evidence for the use of stem cell transplantation in the treatment for acute vascular injury.


Our study took advantage of an animal model suffering from injuries to both the endothelium and the outer membrane of the common carotid artery. When compared with either injury alone, this procedure synergistically enhances necrotic signals, thus facilitating the homing of transplanted stem cells. We used SPIO-DiI dual labeling of stem cells to provide histological evidence for stem cell homing. Our results highlight a previously undocumented application of MR-based live tracking that is devoid of interference from complicating factors such as bleeding on the transplantation site or hemosiderosis. This study indicates that noninvasive, live MR imaging allows dynamic monitoring of magnetically labeled stem cells following transplantation and provides a feasible method to evaluate biological behaviors of transplanted cells during cell-based therapy.


The authors thank Dr. Zuxing Kan at The M.D. Anderson Cancer Center for helping in the preparation of this article.