original article - koreamed

submit.radiology.or.kr J Korean Soc Radiol 2013;68(2):175-185 175

INTRODUCTION

Human mesenchymal stem cell (hMSC) is one of minor pop-ulated adult stem cells in bone marrow, which can differentiate into adipocytes, chondrocytes, and osteocytes under controlled in vivo and in vitro environments. In comparison to embryonic stem cell, hMSC has advantages of easy isolation and a lesser de-gree of ethical concern. Based on these advantages, hMSC has

been known to be a promising treatment option for various or-gan diseases and injuries (1, 2).

In order to evaluate the effect of stem cell treatment, it is im-portant to investigate the efficacy of stem cell delivery and to trace the route of stem cell migration. Although MRI does not provide the necessary sensitivity and contrast for imaging at cel-lular level, the endocytosis of magnetic labels allows cells to be tracked and localized (3). The intracellular confinement of su-

Original ArticlepISSN 1738-2637J Korean Soc Radiol 2013;68(2):175-185

Received July 20, 2012; Accepted November 29, 2012Corresponding author: Jeong Hyun Lee, MDDepartment of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Korea.Tel. 82-2-3010-4352 Fax. 82-2-476-4716E-mail: [email protected]

This work was supported by Asan Medical Center Biomedical Research Center (grant number 08-383).

Copyrights © 2013 The Korean Society of Radiology

Purpose: To quantify in vitro labeling efficiency of protamine sulfate (PS) and poly-L-lysine (PLL) for labeling of human mesenchymal stem cells (hMSCs) with super-paramagnetic iron oxide (SPIO) using multi-echo T2* magnetic resonance (MR) im-aging at 4.7 T. Materials and Methods: The hMSCs were incubated with SPIO-PS or SPIO-PLL complexes. Their effects on the cell metabolism and differentiation capability were evaluated, respectively. The decrease of iron concentrations in the labeled cells were assessed immediately, and at 4 d after labeling using multi-echo T2* MR imaging at 4.7 T. The results were compared with those of Prussian blue colorimetry.Results: The hMSCs were labeled more efficiently by SPIO-PLL than SPIO-PS with-out any significant effect on cell metabolism and differentiation capabilities. It was feasible to quantify the iron concentrations in SPIO-agarose-phantoms and in aga-rose mixture with the labeled cells from T2* maps obtained from multi-echo T2* MRI. However, the iron concentration of the labeled cells was significantly higher by T2*-maps than the results of Prussian blue colorimetry. Conclusion: The hMSCs can be effectively labeled with SPIO-PLL complexes more than with SPIO-PS without significant change in cell metabolism and differentia-tion. In vitro quantification of the iron concentrations of the labeled is feasible from multi-echo T2* MRI, but needs further investigation.

Index termsMesenchymal Stem CellCellular Magnetic ResonanceTransfection Agent

Comparison of Superparamagnetic Iron Oxide Labeling Efficiency between Poly-L-Lysine and Protamine Sulfate for Human Mesenchymal Stem Cells: Quantitative Analysis Using Multi-Echo T2* Magnetic Resonance Imaging1

인간간엽줄기세포에서 Poly-L-Lysine과 Protamine Sulfate 간의 Superparamagnetic Iron Oxide 표지 비교: 다중에코 T2* 자기공명영상기법을 이용한 정량분석1

Ji-Yeon Suh, PhD1,2, Jeong Hyun Lee, MD1, Chang Kyung Lee, MS1, Ji Hoon Shin, MD1, Choong Gon Choi, MD1, Jeong Kon Kim, MD1

1Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea2MRI Team, Korea Basic Science Institute, Daejeon, Korea

Comparison of SPIO Labeling Efficiency between Poly-L-Lysine and Protamine Sulfate for Human Mesenchymal Stem Cells

submit.radiology.or.krJ Korean Soc Radiol 2013;68(2):175-185176

media (Mesenchymal stem cell growth medium; Ronza, Balti-more, MD, USA) were incubated in a 100 mm culture dish with 5% CO2 at 37°C. hMSCs were labeled by a commercially avail-able ferumoxides suspension (SPIO; Feridex IV; Berlex Labora-tories, Wayne, NJ, USA; 11.2 mg/mL) with PLL or PS as a trans-fection agent. For comparison, control hMSCs were incubated without SPIO and transfection agents, while the others were with 50 µgFe/mL of SPIO and a different concentration of PLL (1.5 µg/mL and 2 µg/mL) or PS (10 µg/mL and 15 µg/mL), re-spectively.

In our preliminary study, labeling with high molecular PLL (> 300 kDa) shows severe SPIO aggregation around the cells, while low molecular PLL (1-4 kDa) presented significant lower label-ing efficacy (not published). Therefore, moderate molecular weight PLL (4-15 kDa; Sigma, P6516) was used for this experi-ment (Fig. 1).

A stock of PLL solution prepared at 5 mg/mL was diluted to the concentration of 1 mg/mL with sterile distilled water and put into a mixing tube with 2 different concentrations (1.5 µg/mL and 2 µg/mL). SPIO at a concentration of 50 µgFe/mL and serum-free DMEM medium was added continuously and mixed in a rotating shaker for 1 hr in 4°C cold room. The solution con-taining PLL-SPIO complex after adding the equal volume of se-rum containing medium applied to the hMSC culture dish and incubated for 24 hr. After incubation, the cells were washed 3 times with sterile phosphate-buffered saline (PBS).

Commercially available PS solution (Sigma, P4020) was dilut-ed to the concentration of 1 mg/mL with distilled water and sep-arated into a mixing tube with 2 different concentrations (10 µg/

perparamagnetic iron oxide (SPIO) particles induces a high sus-ceptibility difference between the cells and the surrounding me-dium; the cells become detectable on the image as they appear as hyposignals that are much larger than the true cell size (4). The sensitivity of cellular MRI mainly depends on the content and relaxivity of iron loaded in cells. To enhance labeling of hM-SCs with iron, transfection agents such as protamine sulfate (PS) or poly-L-lysine (PLL) are used. For labeling efficiency of PS and PLL, Arbab et al. (5) reported that magnetic labeling of cells with SPIO-PS complexes is comparable or superior to other SPIO-transfection agent complexes, and iron incorporation into cells were similar to quantities incorporated after labeling with SPIO-PLL. However, their results were not further studies by other researchers.

Along with the content and relaxivity of loaded iron, imaging protocol is important for localization and tracking with cellular MRI, which has been widely investigated in vitro and in vivo by many researchers (3, 6-11). To enable future clinical application and accurate assessment of stem cell therapy, it is a important to develop and optimize methods for a qualitative and a quantita-tive analytic tool using MRI in vitro and in vivo. Therefore, the purpose of this study was to assess the feasibility of a quantitative analytic tool based on MRI for quantification of intracellular iron content with a reference standard of colorimetric assay.

MATERIALS AND METHODS

Cell Labeling with SPIO

hMSCs (Ronza, Baltimore, MD, USA) in standard culture

Fig. 1. Photomicrographs of Prussian blue stained labeled hMSCs with different molecular weight PLLs. A, Labeling with low molecular PLL (1-4 kDa, 2 μg/mL), B, moderate molecular weight PLL (4-15 kDa, 2 μg/mL), and C, with high molecular PLL (> 300 kDa, 2 μg/mL). In contrast to low molecular weight (A) and high molecular weight (C) PLLs, moderate molecular weight PLL labels hMSCs effectively without contamination of extracellular SPIO complexes (B) (Prussian blue staining, × 400 magnification). Note.-hMSCs = human mesenchymal stem cells, PLL = poly-L-lysine, SPIO = superparamagnetic iron oxide

A B C

Ji-Yeon Suh, et al


added by 5 μL of propidium iodide (PI) and 5 μL of Annexin-V-FITC and incubated in the dark for 15 min at room temperature. After adding 400 μL of 1X Binding Buffer, the percent of dead cells and cells undergoing apoptosis were determined by flow cytometry. The annexin-V+/PI- was judged as early apoptosis, and annexin-V+/PI+, as late apoptosis. Under the hypothesis that the transition from annexin-V-/PI- to annexin-V+/PI- was due to the labeling process, the percent differences of annexin-V+/PI- between control (no labeling sample) and each SPIO la-beled samples were used for the indicator of the cytotoxicity.

For assessment of ROS, collected cells were resuspended in prewarmed PBS at 1 × 106 cells/100 µL. The intracellular forma-tion of ROS was detected by CM-H2DCFDA (Molecular Probes, Eugene, OR, USA). CM-H2DCFDA is a nonfluorescent agent that forms fluorescent esters when it is reacted with ROS inside cells. CM-H2DCFDA dye was added at a final concentration of 20 μM and the cells were incubated for 30 min at 37°C. The in-tensity of transformed fluorescent indicators, which is propor-tionate to the cytotoxicity, was analyzed by a fluorescent activat-ed cell sorter immediately.

All of the above experiments were repeated in 3 times for sta-tistical analysis.

Cellular Differentiation into Mesenchymal Lineage

To evaluate the capability of cellular differentiation, the la-beled and non-labeled cells at different concentration of PS and PLL were tested immediately and 4 d after labeling. For osteo-genic differentiation, labeled cells with concentrations of 15 µg/mL PS and 2 µg/mL PLL were plated to 6 well plate at a number of 3 × 104 (3.1 × 103/cm2). Unlabeled cells were prepared in the same way, one for negative control without induction of differ-entiation and the other for positive control. After allowing the cells to adhere for 12 h in growth medium, osteogenesis was in-duced by replacing the media with a commercialized osteogenic medium (Ronza, Baltimore, MD, USA), which was replaced with fresh osteogenic medium every 3-4 d until cellular mor-phology changed into cuboidal shape or delamination occurred indicating calcium deposition. Cultured cells were fixed and stained with 40 mM Alizarin red S for 10 min to detect calcium to assay the osteogenesis.

Labeled subsets identical to the above experiment for osteo-genesis were plated to 6 well plate with the different seeding

mL and 15 µg/mL). SPIO at a concentration of 50 µgFe/mL was added into the mixing tube containing serum-free DMEM me-dium was then shaken intermittently by hand for 5 min. After 5 min, an equal volume of serum containing medium was added to the solution, applied to hMSCs culture dish, and then incu-bated for 24 hr. After incubation, the cells were washed twice with sterile PBS containing heparin (10 U/mL) to dissolve the unlabeled extracellular SPIO-PS complexes, and once with nat-ural PBS.

Labeling efficiency was determined by Prussian blue staining immediately, 4 d (second subculture) and 8 d (third subculture) after labeling in all of the samples. After incubation with SPIO and various concentration of PLL or PS, cells were detached with 0.25% trypsin-EDTA and cytosmeared to a slide glass at × 300 g for 8 min. For Prussian blue staining, the attached cells were in-cubated with 10% potassium ferrocyanide in 20% hydrochloric acid and then counter-stained with nuclear fast red for 5 min. Stained slides were reviewed by Olympus microscope at × 400 magnification using DP controller/manager software. Each ex-periment was repeated 3 times.

Cellular Viability, Proliferation and Apoptosis

To assess viability and proliferation of labeled hMSCs, Alamar Blue assay (Biosource, Camarillo, CA, USA) was performed at immediate (at first subculture), 4 d (at second subculture) and 8 d (at third subculture) after labeling for each subset. Alamar Blue is a sodium salt of resazurin, which is a blue dye as an indi-cator for cell viability. Dark blue colored resazurin is reduced to bright pink colored, fluorescent compound in the presence of live cells. The cells growing in 60 mm culture dishes were tryp-sinized and collected. One tenth volume of total cells in 100 µL complete medium was seeded in 96-well plate and then 10 µL Alamar Blue dye was added to each well. The color changes were measured at a wavelength of 570 nm and 600 nm follow-ing incubation for 8 h. Percent reduction of Alamar Blue of each sample was calculated.

The number of cells undergoing apoptosis was determined by Annexin-V-FITC apoptosis detection kit I (BD PharmagenTM, Franklin Lakes, NJ, USA) using fluorescent-activated cell sorter immediately (at first subculture) and 4 d (at second subculture) after cell labeling. The 1 × 106 labeled cells were collected and suspended in 100 μL of 1X Binding Buffer. The suspension was



using SPIO-PS or SPIO-PLL complexes at day 0 were trypsin-ized, collected and resuspended in 750 µL of PBS and mixed with the prewarmed 750 µL of 1% agarose solution in Eppendorf tubes. For MR imaging of the labeled cells at day 4, 1.5 × 107 cells in 500 µL of PBS were collected to the Eppendorf tubes contain-ing prewarmed 500 µL of 1% agarose solution. For each experi-ment, a calibration curves were plotted from 5 different concen-trations of the SPIO agarose phantoms ranging from 0.39 to 6.25 μgFe in 1000 μL mixed with the same number of unlabeled hMSCs. All of the samples were imaged with MRI scanner us-ing the MR parameters described above.

After T2* imaging of the SPIO agarose phantoms and the la-beled cells mixed with agarose within Eppendorf tubes, the im-aging data were transferred to a personal computer to obtain T2* values with a commercial software (Origin® 7.5, OriginLab Corp.), by performing fitting of the 50 echo images to a monoex-ponential decay as a function of echo time. Using the software, one of the authors measured T2* values of each samples by man-ual drawing of region-of-interests on the T2* images of the sam-ples. After obtaining a regression equation from the data of the SPIO phantoms, the iron concentration of the labeled cells were also calculated, which were correlated with the results of Prus-sian blue colorimetric assay.

Statistic Analysis

All of the experiments were independently repeated in 3 times and the results were presented as an average ± standard devia-tion. Values for evaluation of cellular viability (Alamar Blue) and ROS generation were expressed as a percentage compared to unlabeled control and apoptosis detection was showed as the percentage of annexin-V+/PI-. ROS generation and SPIO con-tents per cell by colorimetry were presented as absolute values. To discriminate the significant differences among the groups Kruskal-Wallis were performed. Less than 0.05 of p-value con-sidered to indicate a significant difference.

RESULTS

Cell Labeling

Prussian blue staining of SPIO-labeled hMSCs by PS or PLL demonstrated 100% of labeling efficiencies by both of 2 different concentrations of PLL and PS, respectively. In case of PLL, small

density at a number of 1.92 × 105 (2 × 104/cm2) and cultured 48 hours until the cultures reached confluence. At 100% conflu-ence, adipogenic differentiation was induced by 3 cycles of in-duction and maintenance. Each cycle consisted of 3-day culture with supplemented adipogenesis induction medium (37°C, 5% CO2) and following 2-day culture with supplemented adipogen-ic maintenance medium. The difference of induction medium lied on the addition of dexamethasone, indomethacin and IBMX (Cambrex, Baltimore, MD, USA) to the maintenance medium. After completion of 3 cycles, the cells were cultured for 7 more days in the maintenance medium. Then, cells were fixed and stained with Oil Red O to detect lipid vacuoles to assess the adi-pogenesis.

Iron Concentration by Colorimetric Assay

First, a calibration curve using Prussian blue colorimetric assay was plotted from different concentrations of SPIO ranging from 0.175 to 11.2 µg/200 µL, which yielded an equation using linear regression analysis. To exclude the optical density measured by the cells themselves, not intracellular iron particles, 3 × 105 cells (the same number of cells with test samples) were added to each well. 3 × 105 of cells immediately after labeling suspended in 100 µL of were poured in 96-well plate. Then, the equal amount of 10% of potassium ferrocyanide and 20% of hydrochloric acid were added to the wells. The optical density of the wells was read 15 min later at the 650 nm wavelength. The SPIO concen-trations per each well were calculated from the equation from the calibration curve. Finally, the SPIO concentration for each cell was figured out by dividing the total SPIO concentration by total cell number per well.

Cellular Magnetic Resonance Imaging

MRI was performed with a 4.7-T magnetic resonance (MR) scanner (Biospec; Bruker Biospin, Billerica, MA, USA) using a quadrature volume coil (diameter, 75 mm). The multi-echo gra-dient echo T2* sequence was used for quantitative T2* mapping with 50 echoes from 3.2 msec to 225 msec (repetition time = 5000 msec; flip angle, 30°). Images were obtained with a matrix size of 128 × 128, one measurement, slice thickness of 2 mm, slice number of 3 and field of view of 50 × 50 mm. The scan time was 8 min.

For quantification of intracellular iron, 3.0 × 106 labeled cells

Ji-Yeon Suh, et al


amount of Prussian blue stained SPIO was markedly decreased by repeated subcultures (at day 8) within the hMSCs labeled ei-ther by SPIO-PS or SPIO-PLL complexes.

amount of SPIO-PLL complexes proportionate to the PLL con-centration were observed around the cells, which completely disappeared at the second subculture (Fig. 2). However, the

Fig. 2. Photomicrographs after Prussian blue staining of SPIO-labeled hMSCs. The photographs immediately after labeling (at the 1st subculture) were located at the upper row, while those 4 days (at the 2nd subculture) and 8 days (at the 3rd subculture) after labeling at the middle and bot-tom rows, respectively (A, F, K: control; B, G, L: 10 µg/mL of PS; C, H, M: 15 µg/mL of PS; D, I, N: 1.5 µg/mL of PLL; E, J, O: 2 µg/mL of PLL). Note that significant decrease in the amount of stained SPIOs within labeled cells and the increase in the number of unlabeled cells in either cas-es by PS or PLL with repeated subcultures (Prussian blue staining, × 400). Note.-hMSCs = human mesenchymal stem cells, PLL = poly-L-lysine, PS = protamine sulfate, SPIO = superparamagnetic iron oxide

K

F

A

L

G

B

M

H

C

N

I

D

O

J

E

Fig. 4. Graph shows the relative production of ROS in the labeled cells expressed as means of the average percentage of ROS in the control cells immediately (black) and 4 days (white) after labeling. Data were derived from three independent experiments. PS10, 10 μg/mL of PS; PS15, 15 μg/mL of PS; PLL1.5, 1.5 μg/mL of PLL; PLL2.0, 2.0 μg/mL of PLL.Note.-PLL = poly-L-lysine, PS = protamine sulfate, ROS = reactive ox-ygen species

0

Fig. 3. Bar graph of relative rates per control determined by Alamar Blue assay immediately (white), 4 days (gray) and 8 days (black) after labeling. As shown in this figure, the viability immediately after label-ing and the proliferating ability after two repeated subcultures were not significantly changed from the controls (p > 0.05). PS10, 10 μg/mL of PS; PS15, 15 μg/mL of PS; PLL1.5, 1.5 μg/mL of PLL; PLL2.0, 2.0 μg/mL of PLL. Note.-PLL = poly-L-lysine, PS = protamine sulfate

0

40

20

20

80

6040

120

10060

160

140

180

80

200100

Mea

n ge

omet

ric fl

uore

scen

ce

% R

educ

ed (c

ontr

ol re

lativ

e)

ControlControl PS10PS10 PS15PS15 PLL1.5PLL1.5 PLL2PLL2

d0 d4 d0 d4 d8



Cellular Viability, Proliferation and Apoptosis

The viability and proliferating ability of SPIO-PS/PLL labeled cells at a different concentration expressed as a percentage of control unlabeled cells by Alamar Blue assay. The results of 3 in-dependent experiments were presented in Fig. 3.

The generation of ROS in the labeled cells was expressed as a percentage of the control cells immediately and 4-d after label-ing (Fig. 4). Although there was a tendency toward higher ROS production in the labeled hMSCs, there was no significant in-crease in ROS production compared with the control immedi-ately and 4-d after labeling.

The results of flow cytometry of the proportion of apoptotic and dead cells in the labeled and non-labeled hMSCs immedi-ately and 4-d after labeling are shown in Figs. 5 and 6. There is no significant increase in the number of apoptotic or dead cells as compared with the number of apoptotic or dead control cells at the different time points.

Cellular Differentiation into Mesenchymal Lineage

The capacity of hMSCs to differentiate into osteogenic and adipogenic cells were determined by histologic examination af-ter induction of differentiation. There was no inhibition to os-teogenic and adipogenic differentiation when labeled by SPIO-PS (50 µg/mL of SPIO; 15 µg/mL of PS) or SPIO-PLL (50 µg/mL of SPIO; 2.0 µg/mL of PLL) complexes (Figs. 7, 8). The capacity to differentiate into osteogenic and adipogenic cells were pre-served not only immediately (at the first subculture), but 4 d (at the second subculture) after labeling.

Iron Concentration by In Vitro Colorimetric Assay

The iron contents per the labeled cells were significantly differ-ent according to the transfection agents used for labeling. The iron contents labeled by PLL were significantly higher than those by PS. In case of concentration, there was no significant difference in the iron contents according to the 2 different concentrations of PS (1.97 ± 0.476 pg/cell vs. 2.31 ± 0.536 pg/cell, p-value > 0.05). However, the iron content was significantly higher for the labeled cells by 2.0 μg/mL of PLL than those by 1.5 μg/mL of PLL (8.22 ± 1.443 pg/cell vs. 5.98 ± 1.552 pg/cell, p-value < 0.05) (Fig. 9).

Cellular MR Imaging

From the calibration curves plotted from the SPIO agarose

Fig. 5. Graphs (A-E) show results of analysis of apoptotic and dead cells in samples of control (i.e., nonlabeled) (A), SPIO-PS complex-la-beled (B, C), and SPIO-PLL complex-labeled (D, E) hMSCs, conducted with fluorescent-activated cell sorting. There is no significant increase in the number of apoptotic or dead cells as compared with the number of apoptotic or dead cells in the control. PS10, 10 μg/mL of PS; PS15, 15 μg/mL of PS; PLL1.5, 1.5 μg/mL of PLL; PLL2.0, 2.0 μg/mL of PLL.Note.-hMSCs = human mesenchymal stem cells, PLL = poly-L-lysine, PS = protamine sulfate, SPIO = superparamagnetic iron oxide

Fig. 6. Graph shows the rates of apoptotic cells per total number of cells at the right lower quadrant in the dot graphs immediately (black) and 4 days (white) after labeling. There was no significant difference in the apoptosis rates among the five groups immediate and 4 days after labeling. PS10, 10 μg/mL of PS; PS15, 15 μg/mL of PS; PLL1.5, 1.5 μg/mL of PLL; PLL2.0, 2.0 μg/mL of PLL. Note.-PLL = poly-L-lysine, PS = protamine sulfate

0.0

0.2

0.6

0.4

0.8

1.0

1.2

1.4

Perc

ent o

f ann

exin

-V+/

PI-

cell

Control PS10 PS15 PLL1.5 PLL2

d0 d4

Prop

idiu

m lo

dide

Prop

idiu

m lo

dide

Prop

idiu

m lo

dide

Annexin-V Annexin-V

Control

PS10

PLL1.5

PS15

PLL2.0

Annexin-V

Annexin-V

Annexin-V

Prop

idiu

m lo

dide

Prop

idiu

m lo

dide

A

B C

D E

Ji-Yeon Suh, et al


beled cells by 2 different concentrations of PS or PLL were also estimated. As shown in Fig. 11, the iron concentrations per each cell were significantly overestimated compared to the results of Prussian blue colorimetric assay. However, the decrease of the intracellular iron concentration after 4 d of subculture was clear-ly demonstrated in all of the samples of PS and PLL, which is

phantoms mixed with unlabeled cells yielded equations by lin-ear regression analysis for each time point as follows (Fig. 10);

Day 0: T2* = 14.12069 + 68.12794 × exp(-1.26348Fe)Day 4: T2* = 9.49466 + 23.72912 × exp(-0.75032Fe)

where exp is exponential, and Fe, iron concentration.From the above equations, the iron concentrations of the la-

Fig. 7. Photomicrographs showing the results of osteogenic induction of hMSCs immediately (A-D) and 4-days (E-H) after labeling by the SPIO-PS (C, G) and the SPIO-PLL (D, H) complexes. A and E are negative and B and F are positive controls, respectively. Osteogenesis is demon-strated with Alizarin red S stain by detection of calcium or calcium salts. Note calcium derivatives stained red and microscopic alteration in cell morphology, gap formation and delamination, as they differentiate and mineralize (Alizarin red S stain, × 100).Note.-hMSCs = human mesenchymal stem cells, PLL = poly-L-lysine, PS = protamine sulfate, SPIO = superparamagnetic iron oxide

Fig. 8. Photomicrographs showing the results of adipogenic induction of hMSCs immediately (A-D) and 4-days (E-H) after labeling by the SPIO-PS (C, G) and the SPIO-PLL (D, H) complexes. A and E are negative and B and F are positive controls, respectively. Oil red O stain for lipid vacuoles and Prussian blue stain for SPIO particles were co-performed to demonstrate adipogenesis induction and iron deposition. Note the lipid vacuoles stained red and the morphological change from spindle to cuboial in shape suggesting adipocyte differentiation of the hMSCs. Prussian blue stained SPIOs are markedly decreased within the adipogenic cells after 2nd subculture (G, H) (Oil red O and Prussian blue staining, × 100). Note.-hMSCs = human mesenchymal stem cells, PLL = poly-L-lysine, PS = protamine sulfate, SPIO = superparamagnetic iron oxide

E

E

A

A

F

F

B

B

G

G

C

C

H

H

D

D



demonstrated in the labeled cells by SPIO-PS complexes (1.5-1.9 pg/mL vs. 26-29 pg/mL). The major cause of the discrepancy be-tween the colorimetry and multi-echo T2* MR could be led from the fact that T2*-weighted images are vulnerable from air-tissue interface in the agarose phantom, which causes signal loss confounding the negative contrast enhancement effect by SPIO agent (15) and needs further investigation. For in vitro measure-ment of the intracellular iron concentration with multi-echo T2* MR, sample preparation technique should be further inves-tigated to reduce susceptibility artifact. However, this can be ap-plied only for in vitro situation but not for in vivo state, because

well corresponded to the results of the histologic examination of the labeled cells.

DISCUSSION

Detection and observation of the fate of the transplanted cells in a recipient is important for monitoring and evaluation of the cell therapy effectiveness. For non-invasive in vivo monitoring of the cell therapy, MRI is a promising tool, because it can be utilized to observe the dynamic migration of cells with long im-aging window, high temporal and spatial resolution, and good contrast, without ionic radiation (12). For cellular tracking with MR, labeling of the cells should be preceded with contrast agent such as SPIO, which step is modified by a transfection agent to maximize the labeling efficiency of stem cells.

For better visualization of the cells, the concentration of the incorporated iron particles is very important. Although the iron concentrations can vary according to the labeling protocol, cell type, and transfection agent, the reported iron concentration for labeling of hMSCs using SPIO-PLL complexes is also variable from 5 pg/mL to 64 pg/mL according to the literature (9, 10, 13, 14). From our results of the colorimetry, the iron concentrations were 5.5-7.8 pg/cell by SPIO-PLL complexes, which are in the range of those published in the articles (9, 10, 13, 14). However, the iron concentrations measured by multi-echo T2* MR were significantly overestimated compared to those of the colorime-try, as 45-50 pg/mL by SPIO-PLL complexes. This trend was

Fig. 9. Graph showing the calculated iron concentration of the hM-SCs immediately after labeling from the colorimetric assay. Note.-hMSCs = human mesenchymal stem cells, PLL = poly-L-lysine, PS = protamine sulfate

Fig. 10. Graphs showing the calibration curves from cellular magnetic resonance imaging of the SPIO-agarose phantoms mixed with the same amount of labeled cells by PS or PLL. A is the calibration curve for the determination of iron quantity per cell at immediately after la-beling, 3 × 106 of unlabeled cells were mixed with 1% agarose solu-tion. While 1.5 × 107 of unlabeled cells were added for the calibration curve to examine the iron concentration within the cells at the 4 days after labeling (B). The graphs show significant correlation between the iron concentration of the T2* values. Note.-PLL = poly-L-lysine, PS = protamine sulfate, SPIO = superpara-magnetic iron oxide

0

8

10

10

12

20

14

16

30

18

20

40

22

24

50

26

60

28

2

4

6

8

10

Fe c

once

ntra

tion

(pg/

each

cel

l)

T2* (

2nd

subc

ultu

re)

T2* (

1st s

ubcu

lture

)

PS10

0

0

1

1

2

2

3

3

4

4

5

5

Equation: y = y0 * A*exp (R0*x)Weighting:y No weightingChi*2yDoF = 1.34641R*2 = 0.98762y0 9.49466 ± 1.22179A 23.72912 ± 2.14982R0 -0.75032 ± 0.1686

Equation: y = y0 * A*exp (R0*x)Weighting:y No weightingChi*2yDoF = 22.1013R*2 = 0.95599y0 14.12069 ± 3.76318A 68.12794 ± 14.4201R0 -1.26348 ± 0.44354

6

6

7

7PS15 PLL1.5

Fe conc. (ug/mL)

Fe conc. (ug/mL)PLL2

d0

A

B

T2*

T2*

Ji-Yeon Suh, et al


time might affect the quality of magnetic labeling by PLL: the longer the incubation period, the more excessive formation of macroscopically visible SPIO-PLL complexes not incorporated into the cytoplasms due to their larger size. Our results also demonstrated that the labeling efficiency and the quality were affected by both optimal ratio between SPIO and PLL (50 : 1.5) and molecular weight of PLL. Therefore, magnetic labeling of the hMSCs by PS or PLL should be further investigated for opti-mization of the magnetic labeling of hMSCs for cellular MRI.

There are limitations in our study. The colorimetric assay for quantification of intracellular iron concentration was carried out for intact labeled cells, without any cytolytic process. Considering colorimetric assay is usually performed after cytolysis, this might be a factor of discrepancy in the results between colorimetric as-say and T2* mapping using multi-echo T2* MR imaging.

In conclusion, hMSCs can be effectively labeled with SPIO-PLL complexes more than with SPIO-PS without significant change in cell metabolism and differentiation. Although quanti-fication of the iron concentrations of the labeled cells in agarose mixture was feasible from T2* maps obtained from multi-echo T2* MRI, it has a tendency of significant overestimation com-pared to the results of Prussian blue colorimetry. Therefore, fur-ther investigation is still necessary for T2* MR-based analytic method to be used as a standard analytic tool in near future.

REFERENCES

1.DazziF,HorwoodNJ.Potentialofmesenchymalstemcell

therapy.CurrOpinOncol2007;19:650-655

2.SerakinciN,KeithWN.Therapeuticpotentialofadultstem

cells.EurJCancer2006;42:1243-1246

3.MowatP,FranconiF,ChaponC,LemaireL,DoratJ,HindréF,

etal.EvaluatingSPIO-labelledcellMRefficiencybythree-

dimensionalquantitativeT2*MRI.NMRBiomed2007;20:

21-27

4.BulteJW,DuncanID,FrankJA.Invivomagneticresonance

trackingofmagnetically labeledcellsaftertransplanta-

tion.JCerebBloodFlowMetab2002;22:899-907

5.ArbabAS,YocumGT,KalishH,JordanEK,AndersonSA,

KhakooAY,etal.Efficientmagneticcell labelingwith

protaminesulfatecomplexedtoferumoxidesforcellular

MRI.Blood2004;104:1217-1223

the sample preparation step is omitted for cellular MRI. On the contrary, the visibility of the labeled cells mainly affected by the intracellular iron concentration might be the most important factor for monitoring of the transplanted cells, which should be further investigated with future studies.

Another result of our study is that although both SPIO-PS and SPIO-PLL complexes were able to label the hMSCs effectively with no harmful effect on the cellular viability, proliferation, apop-tosis and differentiation capability, the amount of intracellular in-corporated iron was significantly larger by SPIO-PLL than SPIO-PS complexes, not only by Prussian blue colorimetry but also by T2* maps obtained from multi-echo T2* MR imaging at 4.7 T.

In detail, although all of the hMSCs incorporated blue colored irons within their cytoplasms either by PS or PLL, the degree of staining was more prominent by PLL, which became more prom-inent after the first subculture of the labeled cells. In contrast to PLL, PS showed very clear labeling without background con-tamination by the extracellular SPIO-PLL complexes that may hinder accurate assessment of intracellular incorporated iron content in vivo (5). Considering that iron plays a pivotal role en-abling cellular tracking by MR, higher incorporation of intracel-lular iron by PLL is an important advantage as a transfection agent over PS. Arbab et al. (5) suggested that the incubation

Fig. 11. Graph showing the iron concentrations of the labeled cells by the quantitative T2* magnetic resonance imaging (50-PS10, 10 μg/mL of PS; 50-PS15, 15 μg/mL of PS; 50-PLL1.5, 1.5 μg/mL of PLL; 50-PLL2.0, 2.0 μg/mL of PLL). Gray bars indicate the iron concentrations of the cells immediately after labeling, and white bars those of the cells 4 days after labeling. Although the iron concentrations of the labeled cells were overestimated compared with the results of the colorimetry, those clearly show the relative amount of the intracellular iron be-tween the PS and PLL groups.Note.-PLL = poly-L-lysine, PS = protamine sulfate

0

10

30

20

40

50

60

Fe c

onc.

(pg/

cell)

50-PS10 50-PS15 50-PLL1.5 50-PLL2

d0 d4



DelmasY,etal.MRevaluationoftheglomerularhoming

ofmagneticallylabeledmesenchymalstemcells inarat

modelofnephropathy.Radiology2006;238:200-210

12.UngerEC.Howcansuperparamagnetic ironoxidesbe

usedtomonitordiseaseandtreatment?Radiology2003;

229:615-616

13.FrankJA,MillerBR,ArbabAS,ZywickeHA,JordanEK,

LewisBK,etal.Clinicallyapplicablelabelingofmammali-

anandstemcellsbycombiningsuperparamagneticiron

oxidesandtransfectionagents.Radiology2003;228:480-

487

14.ArbabAS,BashawLA,MillerBR,JordanEK,LewisBK,Ka-

lishH,etal.Characterizationofbiophysicalandmetabolic

propertiesofcells labeledwithsuperparamagnetic iron

oxidenanoparticlesandtransfectionagentforcellularMR

imaging.Radiology2003;229:838-846

15.BosC,DelmasY,DesmoulièreA,SolanillaA,HaugerO,Gros-

setC,etal.InvivoMRimagingofintravascularlyinjected

magneticallylabeledmesenchymalstemcellsinratkidney

andliver.Radiology2004;233:781-789

6.WalczakP,KedziorekDA,GiladAA,BarnettBP,BulteJW.

ApplicabilityandlimitationsofMRtrackingofneuralstem

cellswithasymmetriccelldivisionandrapidturnover:the

caseoftheshivererdysmyelinatedmousebrain.MagnRe-

sonMed2007;58:261-269

7.OudeEngberinkRD,vanderPolSM,DöppEA,deVriesHE,

BlezerEL.ComparisonofSPIOandUSPIOforinvitrola-

belingofhumanmonocytes:MRdetectionandcellfunc-

tion.Radiology2007;243:467-474

8.OppitzM,PintaskeJ,KehlbachR,SchickF,SchriekG,

BuschC.Magneticresonanceimagingof iron-oxidela-

beledSK-Mel28humanmelanomacellsinthechickem-

bryousingaclinicalwholebodyMRIscanner.MAGMA

2007;20:1-9

9.JuS,TengGJ,LuH,ZhangY,ZhangA,ChenF,etal.Invivo

MRtrackingofmesenchymalstemcellsinratliverafter

intrasplenictransplantation.Radiology2007;245:206-215

10.JuS,TengG,ZhangY,MaM,ChenF,NiY.Invitrolabeling

andMRIofmesenchymalstemcellsfromhumanumbili-

calcordblood.MagnResonImaging2006;24:611-617

11.HaugerO,FrostEE,vanHeeswijkR,DeminièreC,XueR,

Ji-Yeon Suh, et al


인간간엽줄기세포에서 Poly-L-Lysine과 Protamine Sulfate 간의 Superparamagnetic Iron Oxide 표지 비교: 다중에코 T2*

자기공명영상기법을 이용한 정량분석1

서지연1,2 · 이정현1 · 이창경1 · 신지훈1 · 최충곤1 · 김정곤1

목적: 인간간엽줄기세포(human mesenchymal stem cells; 이하 hMSCs)를 초상자성철산화물(superparamagnetic iron

oxide; 이하 SPIO)로 표지하는 데 있어 protamine sulfate (PS)와 poly-L-lysine (PLL) 간의 표지 효율을 4.7 T magnet-

ic resonance (MR)에서 촬영한 다중에코 T2* MR을 이용하여 정량분석하였다.

대상과 방법: hMSC를 SPIO-PS, SPIO-PLL 복합체로 배양하여 각각의 복합체가 세포에 미치는 영향을 평가하고, 시간

이 흐름에 따라 표지된 세포 내에서 철 함량의 변화를 표지 직후와 표지 후 4일째 4.7 T MR로 촬영한 다중에코 T2*

MR로 얻은 T2* 지도로 분석하였다. Prussian blue 비색법을 최적표준법으로 삼아 철 함량 변화를 비교하였다.

결과: 각각의 복합체는 배양 후 세포 대사와 분화능에 유의한 차이를 보이지 않았으나 표지 효율은 SPIO-PLL 복합체에

서 유의하게 높았다. 다중에코 T2* MR에서 얻은 T2* 지도를 이용하여 SPIO 아가로즈 팬텀과 표지세포를 포함한 아가

로즈 혼합물에서 철의 농도를 구할 수 있었으나 표지된 세포내의 철 농도는 Prussian blue 비색법의 결과에 비해 유의하게

높았다.

결론: hMSC의 SPIO 표지 효율은 SPIO-PS보다 SPIO-PLL 복합체에서 유의하게 높다. 비록 다중에코 T2* MR에서 얻

은 T2* 지도를 이용하여 표지세포의 철 농도를 정량화할 수는 있으나 비색법을 이용한 방법에 비해 유의하게 높았으며

이는 향후 연구를 통해 개선되어야 하겠다.

1울산대학교 의과대학 서울아산병원 영상의학과, 2한국기초과학지원연구원 자기공명영상팀

original article - koreamed

Documents