Using Manganese Oxide Nanoparticles

COMMUNICATIONS Magnetic Resonance in Medicine 60:1–7 (2008)

MR Tracking of Transplanted Cells With “Positive Contrast” Using Manganese Oxide Nanoparticles

Assaf A. Gilad,1,2 Piotr Walczak,1,2 Michael T. McMahon,1,3 Hyon Bin Na,4,5 Jung Hee Lee,6 Kwangjin An,4,5 Taegwhan Hyeon,4,5 Peter C.M. van Zijl,1,3 and Jeff W.M. Bulte1,2,7,8*

Rat glioma cells were labeled using electroporation with either oxide particles (2), or by dextran-coated SPIOs conjugated manganese oxide (MnO) or superparamagnetic iron oxide to monoclonal antibodies (3). Over time, techniques were (SPIO) nanoparticles. The viability and proliferation of SPIO- developed to make the labeling procedure more straight- labeled cells (1.9 mg Fe/ml) or cells electroporated with a low forward and widely applicable, independent of cell type ␮ dose of MnO (100 g Mn/ml) was not significantly different from (4), including the use of dendrimers (5), the human immu- unlabeled cells; a higher MnO dose (785 ␮g Mn/ml) was found to nodeficiency virus (HIV) tat peptide (6), transfection be toxic. The cellular ion content was 0.1–0.3 pg Mn/cell and 4.4 pg Fe/cell, respectively, with cellular relaxivities of 2.5–4.8 s–1 agents (7), and micron-sized particles (8). Recently, elec- –1 troporation (9,10), in which the cell membrane is perme- (R1) and 45–84 s (R2) for MnO-labeled cells. Labeled cells (SPIO and low-dose MnO) were each transplanted in contralat- abilized by a short electrical pulse, was demonstrated to be eral brain hemispheres of rats and imaged in vivo at 9.4T. While efficient in labeling cells with SPIO. SPIO-labeled cells produced a strong “negative contrast” due MRI cell tracking using SPIOs has now been widely used to the increase in R2, MnO-labeled cells produced “positive for tracking transplanted cells in various organs, and has contrast” with an increased R1. Simultaneous imaging of both recently also entered the clinic (11). However, in many transplants with opposite contrast offers a method for MR cases, it is difficult to distinguish labeled cells from other “double labeling” of different cell populations. Magn Reson hypointense regions on T2/T*2-weighted MR images. These Med 60:1–7, 2008. © 2008 Wiley-Liss, Inc. hypointensities can have a physiological origin, such as Key words: manganese oxide; iron oxide; cellular imaging; con- hemoglobin in blood, or a pathological origin, such as trast agent; transplantation; nanoparticles blood clots or experimental, traumatic procedures (e.g., caused by cell injections). One attempt to differentiate Cells for transplantation have been labeled with MR con- iron-labeled cells from blood vessels was to alter the in- trast agents since the beginning of the 1990s. Superpara- haled oxygen levels to reduce the BOLD effect (12). Nev- magnetic iron oxide (SPIO) nanoparticles have been most ertheless, hypointensities on MR images remain a major widely used due to their strong signal attenuation proper- obstacle in the attempt to increase the specificity of cell ties. Initially, cells were labeled with SPIO particles con- tracking, preventing this method from being used in cer- jugated to lectins (1), with a viral envelope containing iron tain applications, particularly those that involve trauma and hemorrhage. Therefore, an alternative method of de- tecting cells is to use contrast agents that create an oppo- 1Russell H. Morgan Department of Radiology, Division of MR Research, The site contrast. Johns Hopkins University School of Medicine, Baltimore, Maryland, USA. Other than iron oxide labels, gadolinium complexes 2Institute for Cell Engineering, Cell Imaging Section The Johns Hopkins Uni- versity School of Medicine, Baltimore, Maryland, USA. have been suggested for labeling fibroblasts with direct 3F.M. Kirby Center for Functional Brain Imaging, Kennedy Krieger Institute, gadolinium diethylenetriamine pentaacetic acid (Gd- Baltimore, Maryland, USA. DTPA) bovine serum albumin (BSA) incubation (13), or 4National Creative Research Initiative Center for Oxide Nanocrystalline Mate- incubation with Gd-DTPA/fatty acid complex (14). In rials, Seoul National University, Seoul, Republic of Korea. 5School of Chemical and Biological Engineering, Seoul National University, these cases, “positive” contrast can be observed. On the Seoul, Republic of Korea. other hand, when internalized into cells, gadolinium- 6 Department of Radiology, Samsung Medical Center, Sungkyunkwan Univer- based agents may exhibit reduced T1 relaxivity as com- sity School of Medicine, Seoul, Republic of Korea. pared to their unbound counterparts in solution. There- 7Department of Chemical and Biomolecular Engineering, The Johns Hopkins University Whiting School of Engineering, Baltimore, Maryland, USA. fore, in certain cases, T2-weighted imaging and hypoin- 8Department of Biomedical Engineering, The Johns Hopkins University tense, magnetic susceptibility-based cell detection may School of Medicine, Baltimore, Maryland, USA. still be preferred when using paramagnetic contrast agents Grant sponsor: National Institutes of Health (NIH); Grant numbers: RO1 (15). NS045062, R21 EB005252; K01 EB006394. Next to gadolinium, manganese is probably the second *Correspondence to: Jeff W.M. Bulte, PhD, The Johns Hopkins University School of Medicine, Russell H. Morgan Department of Radiology and Radio- most used “positive” T1 contrast agent. Increasingly, man- logical Science, Division of MR Research, 217 Traylor Bldg, 720 Rutland Ave, ganese-enhanced MRI is used for anatomical MRI, for the Baltimore, MD 21205. E-mail: [email protected] study of neuronal activity, to monitor neuronal tracts (16), Received 14 May 2007; revised 21 December 2007; accepted 14 February 2008. to study neuronal connectivity in animal models of dis- DOI 10.1002/mrm.21622 ease (17), and to monitor liposomal drug delivery (18,19). Published online in Wiley InterScience (www.interscience.wiley.com). As for MRI cell tracking, manganese salt (MnCl2) has been © 2008 Wiley-Liss, Inc. 1 2 Gilad et al.

used for efficient labeling of lymphocytes in vitro, and a 96-well plates (5 ϫ 103 cells per well). These cells were decrease in T1 was observed for the first 24 h (20). Re- assayed again after 24 h (i.e., a total of 48 h after electro- cently, some of us demonstrated that manganese oxide poration). (MnO) nanoparticles can be used as T1 MR contrast agent For assessment of cell viability, a Calcein-acetyoxy- for various body organs, depicting fine anatomic structures methyl (AM) enzyme assay was used (4892-010-K; Trevi- (21). Here, functionalized MnO nanoparticles prepared by gen Inc.). This assay is based on hydrolysis of Calcein-AM conjugation with a tumor specific antibody were also used by intracellular esterases that produce calcein only in vi- for selectively imaging breast cancer cells in the metastatic able cells. Cells were washed once with 100 ␮l of Cal- brain tumor. cein-AM buffer, and 100 ␮l of Calcein-AM solution was In this study, we evaluated the potential of using these added. Cells were incubated for 30 min at 37°C in a hu- MnO nanoparticles as a novel T1 contrast agent for cell midified 5% CO2 atmosphere. The fluorescence was re- labeling. To this end, we have compared MnO-labeled corded using a 490-nm excitation filter and a 520-nm emis- cells to cells labeled with the SPIO formulation Feridex. sion filter, with the fluorescence intensity being propor- We report here that MnO can be successfully used to tional to the number of viable cells. detect cells with positive contrast in vivo, and that two cell For assessment of proliferation, a MTS ([3-(4,5- populations, one labeled with MnO and the other one with dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4- SPIO, can be detected simultaneously with opposite cell sulfophenyl)-2H-tetrazolium, inner salt) was used (Cell contrast. Titer 96 Aqueous, G3582; Promega). The assay is based on mitochondrial assimilation and conversion of sub- strate. A total of 20 ␮l of Cell Titer 96 Aqueous One MATERIALS AND METHODS Solution Reagent was added per well. Cells were incu- Cell Culture and Labeling bated for2hat37°C in a humidified 5% CO2 atmosphere. The rat glioma cell line 9L was cultured in Roswell Park The absorbance was recorded at 490 nm using a 96-well Memorial Institute (RPMI) medium, supplemented with plate reader. For additional study of cell proliferation, 10% ml fetal bovine serum (FBS) (Sigma), 2 mM L-glu- CellTiter-Blue , Cell Viability Assay (G8080; Promega) tamine (Invitrogen), 1 mM sodium pyruvate, 1:100 ampho- was used. The assay is based on the ability of living cells tericin B solution (Sigma), 0.06 mg/ml gentamicin (QBI), to convert a redox dye (resazurin) into a fluorescent end ␮ and 175 nM 2-mercapto-ethanol (Sigma). product (resorufin): 20 l of Cell Titer Blue reagent was Electroporation (9,10) was used to label cells with MnO added per well. The cells were incubated for2hat37°C in and SPIO particles. MnO nanoparticles were prepared as a humidified, 5% CO2 atmosphere. The fluorescence was described (21), with the nanoparticles having an average recorded using a 96-well plate reader (excitation at 560 nm diameter of 7 nm. Cells were cultured in 80-cm2 flasks and emission at 590 nm). Results are expressed from two ϭ overnight to 80% to 90% confluence. The next day, cells independent experiments (N 6) for MTS and Calcein- ϭ were suspended using trypsin-ethylene diamine tetraace- AM, and from one experiment for Cell Titer Blue (N 3). tic acid (EDTA), washed with phosphate buffered saline (PBS) (without Ca2ϩ and Mg2ϩ), and counted. Cells were MR Phantom Preparation resuspended in PBS and transferred to sterile 0.4-cm gap electroporation cuvettes (Gene Pulser; Bio-Rad); each cu- At 24 h after electroporation, cells were washed twice with vette contained 2 ϫ 106 cells in 580 ␮l of PBS. Metal oxide PBS, harvested using trypsin, and counted. For the cell ϫ 6 ␮ nanoparticles were added at 1.9 mg Fe/ml for SPIO (Feri- pellet phantom, 3.15 10 cells suspended in 100 l PBS dex; Berlex Imaging) and 100 or 785 ␮g Mn/ml for MnO were transferred to 0.2 ml polypropylene tubes (VWR In- nanoparticles. Unlabeled cells that were treated with elec- ternational) and centrifuged for 6 min at 1200 rpm. The troporation without contrast agent were used as controls; supernatant was aspirated and cells were resuspended in ␮ ϫ 6 cells that were incubated with MnO but not electroporated 20 l of PBS. For the gelatin phantom, 4 10 cells ␮ were also included. PBS was added to each cuvette to a suspended in 50 l PBS were transferred to 0.2 mL ␮ final volume of 700 ␮l. The cuvettes were kept on ice for polypropylene tubes and mixed with 100 l of 6% gelatin ϫ 4 1 min, and then electroporated using a BTX electropora- in PBS. Thus, the final cell concentration was 2.7 10 ␮ tion system (ECM830; Harvard Apparatus). The following cells/ l in 4% gelatin. Control samples consisted of PBS conditions were used: pulse strength ϭ 100 V; N pulses ϭ and 4% gelatin in PBS. 5; pulse duration ϭ 5 ms; and pulse interval ϭ 100 ms. After 30 s, cells were transferred to ice for 2 min, sus- Animal Studies pended in culture medium, and transferred to 10-cm culture dishes. Animal experiments were performed in accordance with protocols approved by our institutional Animal Care and Use Committee. At 24 h after electroporation, cells (la- In Vitro Toxicity Tests beled with 1.9 mg Fe/ml for SPIO, 100 ␮g Mn/ml for MnO After electroporation, 5 ϫ 103 cells out of the 2 ϫ 106 cells nanoparticles or unlabeled) were washed twice with PBS, were transferred in triplicate to 96-well plates, with the suspended using trypsin-EDTA, and counted. The cells remaining cells plated into the culture dishes. After 24 h, were centrifuged at 1000 rpm for 5 min and diluted to the the cells in the 96-well plates were assayed for viability. At appropriate concentration. Male Fisher rats (weight 250– the same time, cells in the 10-cm culture dish were washed 350 g) were anesthetized with ketamine/acepromazine with PBS, suspended using trypsin-EDTA, and replated in (100/5 mg/kg) and positioned in a stereotaxic device MR Cell Tracking Using MnO Nanoparticles 3

(Stoelting Lab Standard). A small midline skin incision and 0.25 ␮g of Mn. The cellular iron concentration was was made to expose the skull. Using a 10-␮l Hamilton determined by dividing the total metal content of each syringe with an attached 31G metal needle (Hamilton Co.), sample with the number of cells. Cellular relaxivity was 2 ϫ 105 cells in 2 ␮l PBS each were injected bilaterally into determined using the slope of relaxation enhancement per each striatum (anterior-posterior [AP] ϭ 0.0 medial-lateral unit concentration of intracellular manganese. The relax- ϭ ϭ ϭ [ML] 3.0 dorsal-ventral [DV] 5.0). Cells were injected ivities were calculated as follows: r1 (R1MnO – R1Control)/ ϭ slowly over 4 min, and the needle was left in place for [Mn], and r2 (R2MnO – R2Control)/[MnO]. 1 min before being withdrawn. The incision was sutured and postoperative analgesia was provided (ketoprofen, RESULTS 2 mg/kg). Rats were anesthetized with 1.5% to 2% isoflu- rane and imaged at 24 h (N ϭ 7; five with MnO- and To compare the two contrast agents, 9L cells were labeled SPIO-labeled cells, and three with SPIO- and unlabeled with MnO, SPIO, or not labeled (but still electroporated as cells), at 48 h (N ϭ 1), and at 72 h (N ϭ 1) after cell control). A total of 2 ϫ 105 cells was inoculated in the transplantation. ipsilateral and contralateral hemisphere of Fisher rats. Fig- ure 1a and b show representative images of both SPIO- and MRI and Analysis MnO-labeled cells in the same imaging plane. While the SPIO-labeled cells appear as a hypointense region (dark MR images were acquired on a Bruker 9.4T horizontal bore spot or negative contrast), the MnO-labeled cells appear as magnet equipped with a 30-mm Sawtooth resonator a hyperintense region (bright spot). The hyperintense re- (Bruker) using a multispin multiecho (MSME) pulse se- gion was found to be induced by the MnO and not a result quence. of edema following cell injection, as hemispheres that were injected with unlabeled cells did not show any con- In Vitro Cell Pellet Phantoms trast (Fig. 1c). MnO-labeled cells could be best detected on R1 maps For T1 measurements, the following parameters were used: (Fig. 1d–f). In contrast, SPIO-labeled cells were seen most ϭ ϭ TE 14.1 ms, and TRs 0.2, 0.3, 0.5, 0.8, 1, 1.5, 2, 4, 6, clearly on the R maps (Fig. 1g–i), while MnO-labeled cells ϭ ϭ 2 and 10 s. For T2 measurements: TR 2 s, and TEs 14, 20, were barely visible on the R map. Figure 1j–l shows the ϫ 2 30, 40, 50, and 60 ms. The field of view was 26 26 mm merged images of the R and R maps, demonstrating a ϫ 1 2 with matrix size of 64 64 pixels and a 1.0-mm slice clear proof-of-principle that opposite contrast can be ob- thickness. tained simultaneously using differentially-labeled cell populations. In Vitro Gelatin Phantoms The main advantage of “positive contrast labeling” would be the ability to distinguish cells from blood/hemo- For T measurements, the following parameters were used: 1 siderin-associated hypointense regions. Although some TE ϭ 14.1 ms, and TRs ϭ 0.2, 0.3, 0.5, 0.8, 1, 1.5, 2, 4, 6, pathologies, such as tumor and stroke-related edema, can and 10 s. For T measurements: TR ϭ 2 s, and TEs ϭ 15, 30, 2 appear as hyperintense regions on T -weighted and spin- 45, 60, 75, 90, 105, and 120 ms. The field of view was 27 ϫ 1 density/T -weighted images, they can be easily distin- 18 mm with a matrix size of 128 ϫ 64 and a 1.0-mm slice 2 guished from labeled cells on R maps: the former results thickness. The R and R were calculated by fitting of the 1 1 2 from an increase in tissue water content, which is charac- data to Ln(1 – [I/I ]) ϭ –R*TR and Ln(I) ϭ –R*TE ϩ A, 0 1 2 terized by shortening of R , while MnO-labeled cells result respectively. 1 in an increase of R1. A histogram of the pixel distribution (Fig. 2) reveals that the pixels from the MnO-labeled trans- In Vivo Animal Studies planted cells have a higher R1 but not R2, whereas the ϭ pixels from the brain tissue surrounding the transplanted For R1 maps, the following parameters were used: TE 14.1 ms, and TRs ϭ 0.2, 0.3, 0.5, 0.8, 1, 1.5, 2, 3, 4, and 6 s. cells have R1 values similar to brain and to unlabeled ϭ ϭ transplanted cells. After 2–3 days, it became difficult to For the R2 maps: TR 2 s, and TEs 14, 20, 30, 40, 50, and 60 ms. The field of view was 30 ϫ 30 mm and the matrix detect the MnO-labeled cells, although they were detect- size was 128 ϫ 128 and 64 ϫ 64, for anatomical images and able on the R1 maps (Fig. 3). The reduction in the R1 values could possibly be related to dilution of the contrast agent R1/R2 maps, respectively. The slice thickness was 0.8 mm. following multiple cell divisions. R1 and R2 maps were generated by pixel-by-pixel fitting To characterize the relaxation properties of the labeled using Matlab 6.5.1 (The MathWorks, Inc., USA). In vivo R1 ϭ –TR*R1 cells in vitro, the T and T relaxation for cell pellets and and R2 maps were generated by fitting to I I0*(1–e ) 1 2 ϭ ϩ cells suspended in 4% gelatin were measured 24 h after and Ln(I) –R*2TE A, respectively. electroporation (Fig. 4). The loss of signal from SPIO- labeled cells was not beyond the background noise, thus Measurement of Intracellular Manganese and Iron Content preventing accurate measurements. For cell pellets, the R and Cellular Relaxivity 1 was 0.80 s–1 for unlabeled cells and 4.92 s–1 for cells After the in vitro phantom imaging, all samples were kept labeled with MnO. For the cell suspensions embedded in –1 frozen at –20°C. Samples were assayed for Mn and Fe gelatin, the R1 increased from 0.34 s for unlabeled cells using inductively coupled plasma (ICP) analysis (Varian and 0.41 s–1 for gelatin only to 0.50 s–1 for cells labeled –1 –1 Vista Pro ICP; Microbac Laboratories, Inc., Baltimore, MD, with MnO. The R2 increased from 21 s and 4.8 s for USA). The sample detection threshold was 2.5 ␮gofFe unlabeled cells to 94 s–1 and 7.7 s–1 for MnO-labeled cells 4 Gilad et al.

FIG. 1. In vivo MRI of labeled 9L cells 24 h after transplantation in the striata of rat brain. a–c: Spin echo image (TR ϭ 1000 ms, TE ϭ 14.1 ms). d–f: R1 maps. g–i: R2 maps. j–l: R1/R2 merged maps. Shown are representative images of three out of seven rats, two injected with MnO- and FeO-labeled cells (a,b), and one with FeO-labeled cells and unlabeled cells (c; control). Note the simultaneous double contrast in (j–l).

in phantoms of cell pellets and cells suspended in gelatin, samples or 4% gelatin samples. The iron content in the respectively. Even though in vitro labeled cells exhibit SPIO-labeled cell pellet was 4.4 pg Fe/cell. Iron was not both the R1 and R2 effect, in vivo the R1 effect was more detected in any of the other samples. Thus, the R1 was 4.77 –1 –1 –1 –1 –1 –1 pronounced (Fig. 1). s mM and 2.49 s mM and the R2 was 84.53 s mM For all phantoms, the manganese and iron content was and 45.18 s–1mM–1 for cell pellets and cells suspended in measured with ICP. The Mn concentration was 47.5 ␮g gelatin, respectively. Mn/ml for the phantom with the MnO-labeled cell pellet, To assess any potential toxic effects of MnO labeling which corresponds to 0.3 pg Mn/cell. For the phantom using electroporation, ﬁve independent cell labeling ex- with MnO-labeled cell suspensions in gelatin, the Mn con- periments were performed. 9L cells were electroporated centration was 3.53 ␮g Mn/ml, which corresponds to 0.13 under conditions identical to those that were used for pg Mn/cell. No Mn was detected in control, unlabeled the in vitro and in vivo MRI. Cell viability was deter- MR Cell Tracking Using MnO Nanoparticles 5

FIG. 3. Spin echo images (TR ϭ 1000 ms, TE ϭ 14.1 ms) (a,b) and R1 maps (c,d) of the same rat brain at 1 (a,c) and 3 (b,d) days after cell grafting. The contrast decreases for both MnO- and SPIO- labeled cells over time. Although it is difﬁcult to see MnO-labeled cells on the spin echo image (b), the cells are still visible on the R1 map (d).

mined by testing the mitochondrial activity (MTS) and cellular enzymatic activity (Calcein-AM and Cell Titer Blue), both of which can be correlated with the number of viable cells. After 24 h, a reduction in the number of viable cells was observed for cells labeled with both high and low MnO

FIG. 2. R1 and R2 histograms. a: The R1 of each pixel is presented for two representative rats. A region of interest (ROI) was se- lected to cover both the transplanted cells and brain tissue. Pixels of unlabeled transplanted cells (green, N ϭ 71 pixels) have more uniform and lower R1 values, while MnO-labeled transplanted cells (red, N ϭ 58 pixels) exhibit a broader distribution, shifting toward higher R1 values. The overlapping pixels are from brain tissue surrounding the labeled transplanted cells; the “tail” can be attributed to pixels with a higher R1 that represent MnO- labeled cells. The average R1 for unlabeled and MnO-labeled cells was 0.52 s–1 and 0.63 s–1, respectively. b: These ﬁndings are consistent with the average distribution of R1 from MnO-labeled FIG. 4. In vitro characterization of relaxation enhancement. The ϭ transplanted cells (red, N 5) and unlabeled transplanted cells relative change in signal I/I0 intensity (mean of region of interest (green, N ϭ 2). c: The R2 distribution is similar for MnO-labeled [ROI]) is plotted for a series of TRs and TEs for (a,b) PBS samples transplanted cells (red, N ϭ 5) and unlabeled transplanted cells ({), 4% gelatin only samples (‚), cells suspended in 4% gelatin (green, N ϭ 2), implying that MnO can be used as dominant T1 labeled with MnO (●), and unlabeled cells (ᮀ); (c,d) MnO-labeled cell positive contrast agent. pellets (●) and unlabeled cell pellets (ᮀ). 6 Gilad et al.

approach generates negative contrast on T2-weighted MR images. Here we explored the use of a “positive” contrast agent that can provide opposite contrast on the same type of generated image, using a newly developed MnO nanoparticle preparation (21). Depending on the specific application, labeling cells with MnO nanoparticles could be a viable alternative for iron oxide labeling. These applications include experimental settings where it is difficult to distinguish iron oxide-labeled cells from blood/hemoglobin-derived hypointense regions, or from tissues and organs that have inherently high concentrations of iron, such as the liver, or certain tumors, such as melanomas. An additional application for using MnO nanoparticles includes tracking two different cell populations simultaneously, where one is labeled with MnO and the other with SPIO nanoparticles. In this study, we have shown this proof-of-principle of MR “double labeling” with opposite contrast (Fig. 1). This method will be most suitable for studies that require tracking of two cell populations that are injected in different locations or for tracking two cell populations that are injected at two different time points. The negative contrast enhancement of SPIO-labeled cells was found to be more pronounced than the positive FIG. 5. Assessment of cell viability and toxicity at 24 h (a) and 48 h cellular contrast induced by the MnO nanoparticles. Aside (b) after electroporation with MnO (low: 100 ␮g Mn/ml) or (high: from the differences in relaxivities and concentrations 785 ␮g Mn/ml) and SPIO (1.9 mg Fe/ml). Cell viability was tested using three different assays: MTS (black bars); Cell Titer Blue (gray used, this contrast could also partially result from using a bars); and Calcein-AM (empty bars). All bars represent the aver- labeling procedure that has been optimized for SPIO label- age Ϯ SD of the percentage relative to control (electroporation ing (9,10), with the MnO-labeling procedure currently be- without contrast agent). ing adapted from it. The rationale was to use a protocol that would provide the maximum labeling for SPIO, serv- ing as a reference for the best labeling, with minimal concentrations (t-test, unpaired, two-tailed, P Ͻ 0.05), but differences between the labeling procedures. It may be not those labeled with SPIO (Fig. 5). Notably, cells that possible that further optimization of the MnO labeling were incubated with MnO without electroporation protocol can increase the concentration of intracellular showed a similar degree of toxicity (Fig. 5a), indicating manganese, along with detection of cells over a longer time that the contrast agent and not the electroporation itself frame. leads to cellular impairment. One of the drawbacks of labeling cells with exogenous To determine whether surviving cells can proliferate at nanoparticles is the loss of signal over time, as shown in the same rate as the nonlabeled cells, 24 h after electropo- Fig. 3. The loss of signal can be attributed to either bio- ration, labeled cells were washed, counted again, and degradation of the contrast agent or, in the case of dividing transferred into new culture plates. The cell viability was cells, dilution of label that is amplified in asymmetric cell tested again after 24 h (total 48 h after electroporation) division (22). One approach to overcome this is to use MRI with MTS, Calcein-AM, and Cell Titer Blue. At this time, reporter genes, in which daughter cells retain a constant no significant difference could be observed between any of amount of the contrast-enhancing agent after each cell the conditions, except for cells that were electroporated division (23,24). with the higher MnO concentration (t-test, unpaired, two- We have employed a prototype MnO nanoparticle that tailed, P Ͻ 0.05). Of note is that cells electroporated with has not yet been fully optimized for biocompatibility SPIO showed an enhanced proliferation, even relative to and biomedical use. We observed a transient toxic effect control cells (Fig. 5b). after 24 h in cells that were electroporated with MnO. The overall findings demonstrate that cells labeled with However, after another 24 h, the surviving cells showed the lower MnO concentration recover 48 h after labeling viability similar to that of controls, as manifested by (at least in vitro) and show viability similar to control mitochondrial and cytoplasmic enzymatic activity. It is cells. well known that Mn can be toxic to mammalian cells. However, it has no cellular toxicity at concentrations Ͻ2 mM, as reported for lymphocytes (20), and at con- DISCUSSION centrations Ͻ0.8 mM, as reported for human fibroblasts In this study, we compared labeling of cells with two and cancer cells (21). In this study, we used 1.85 mM different contrast agents, i.e., MnO and SPIO nanopar- and 14.5 mM Mn to get maximal loading of the cells; ticles. Both are metal oxides; SPIO nanoparticles have while the high dose was found to be toxic, the low dose been in use for several years as an effective T2/T*2 contrast of MnO did not induce significant toxicity, in analogy agent for labeling and tracking cells. This cell tracking with those studies. MR Cell Tracking Using MnO Nanoparticles 7

In vitro measurements of R1 and R2 showed good 6. Josephson L, Tung CH, Moore A, Weissleder R. High-efficiency intra- correlation with the Mn content in labeled cells as mea- cellular magnetic labeling with novel superparamagnetic-Tat peptide sured with ICP. Interestingly, the cellular T and T conjugates. Bioconjug Chem 1999;10:186–191. 1 2 7. Frank JA, Miller BR, Arbab AS, Zywicke HA, Jordan EK, Lewis BK, relaxivities were different for cell pellet and cells sus- Bryant LH Jr, Bulte JW. Clinically applicable labeling of mammalian pended in gelatin, indicating that the microenvironment and stem cells by combining superparamagnetic iron oxides and trans- of the transplanted cells has considerable impact on the fection agents. Radiology 2003;228:480–487. relaxivities. This is also true in vivo, in which the re- 8. Hinds KA, Hill JM, Shapiro EM, Laukkanen MO, Silva AC, Combs CA, laxation measured from each voxel is composed not Varney TR, Balaban RS, Koretsky AP, Dunbar CE. Highly efficient endosomal labeling of progenitor and stem cells with large magnetic only from the transplanted cells but also from a variety particles allows magnetic resonance imaging of single cells. Blood of inhabitant cells and the extracellular matrix. The 2003;102:867–872. mechanism by which the MnO nanoparticles generate 9. Walczak P, Ruiz-Cabello J, Kedziorek DA, Gilad AA, Lin S, Barnett B, contrast is not fully understood. The dipolar relaxation Qin L, Levitsky H, Bulte JW. Magnetoelectroporation: improved that underlies T contrast must arise from direct inter- labeling of neural stem cells and leukocytes for cellular magnetic 1 resonance imaging using a single FDA-approved agent. Nanomedi- actions between magnetic moments at a short spatial cine 2006;2:89–94. scale. Several possibilities arise that can accomplish 10. Walczak P, Kedziorek DA, Gilad AA, Lin S, Bulte JW. Instant MR this. A first explanation would be the presence of Mn on labeling of stem cells using magnetoelectroporation. Magn Reson Med the outer shell of the crystal or Mn leaching out from the 2005;54:769–774. 11. de Vries IJ, Lesterhuis WJ, Barentsz JO, Verdijk P, van Krieken JH, particles. A second would be, if phospholipid-polyeth- Boerman OC, Oyen WJ, Bonenkamp JJ, Boezeman JB, Adema GJ, ylene glycol (PL-PEG) allows water diffusion across the Bulte JW, Scheenen TW, Punt CJ, Heerschap A, Figdor CG. Magnetic coating toward the crystal outer surface, that proton resonance tracking of dendritic cells in melanoma patients for mon- relaxation is enhanced after direct contact with the crys- itoring of cellular therapy. Nat Biotechnol 2005;23:1407–1413. tal and water diffuses outward. Finally, as the phospho- 12. Himmelreich U, Weber R, Ramos-Cabrer P, Wegener S, Kandal K, Shapiro EM, Koretsky AP, Hoehn M. Improved stem cell MR detect- lipid shell is directly in contact with the crystal and the ability in animal models by modification of the inhalation gas. Mol total particle is moving slow, there is a rapid dipolar Imaging 2005;4:104–109. transfer (spin diffusion) between them. The same is true 13. Granot D, Kunz-Schughart LA, Neeman M. Labeling fibroblasts with between the phospholipids and PEG. As PEG contains biotin-BSA-GdDTPA-FAM for tracking of tumor-associated stroma by rapidly exchanging OH protons, the paramagnetic relax- fluorescence and MR imaging. Magn Reson Med 2005;54:789–797. 14. Himmelreich U, Aime S, Hieronymus T, Justicia C, Uggeri F, Zenke M, ation effects may be transferred through exchange to the Hoehn M. A responsive MRI contrast agent to monitor functional cell solvent water. Although we do not have direct evidence status. Neuroimage 2006;32:1142. to support any of these possible mechanisms we are 15. Modo M, Mellodew K, Cash D, Fraser SE, Meade TJ, Price J, Williams aiming to elucidate this in future work. SC. Mapping transplanted stem cell migration after a stroke: a serial, in In summary, the present study exemplifies the use of vivo magnetic resonance imaging study. Neuroimage 2004;21:311–317. 16. Silva AC, Lee JH, Aoki I, Koretsky AP. Manganese-enhanced magnetic MnO and SPIO nanoparticles for obtaining dual contrast of resonance imaging (MEMRI): methodological and practical consider- two different cell populations, which may encourage fur- ations. NMR Biomed 2004;17:532–543. ther investigations into developing and optimizing nano- 17. Pelled G, Bergman H, Ben-Hur T, Goelman G. Manganese-enhanced particles that can provide positive contrast. MRI in a rat model of Parkinson’s disease. J Magn Reson Imaging 2007;26:863–870. 18. Viglianti BL, Abraham SA, Michelich CR, Yarmolenko PS, MacFall JR, ACKNOWLEDGMENTS Bally MB, Dewhirst MW. In vivo monitoring of tissue pharmacokinet- ics of liposome/drug using MRI: illustration of targeted delivery. Magn We thank Danny Shin and Segun M. Bernard for assistance Reson Med 2004;51:1153–1162. with the experiments. Supported by NIH grants RO1 19. Viglianti BL, Ponce AM, Michelich CR, Yu D, Abraham SA, Sanders NS045062 (to J.W.M.B.), R21 EB005252 (to J.W.M.B), and L, Yarmolenko PS, Schroeder T, MacFall JR, Barboriak DP, Colvin OM, Bally MB, Dewhirst MW. Chemodosimetry of in vivo tumor K01 EB006394 (to M.T.M.). liposomal drug concentration using MRI. Magn Reson Med 2006;56: 1011–1018. 20. Aoki I, Takahashi Y, Chuang KH, Silva AC, Igarashi T, Tanaka C, Childs REFERENCES RW, Koretsky AP. Cell labeling for magnetic resonance imaging with 1. Norman AB, Thomas SR, Pratt RG, Lu SY, Norgren RB. Magnetic the T1 agent manganese chloride. NMR Biomed 2006;19:50–59. resonance imaging of neural transplants in rat brain using a superpara- 21. Na HB, Lee JH, An K, Park YI, Park M, Lee IS, Nam DH, Kim ST, Kim magnetic contrast agent. Brain Res 1992;594:279. SH, Kim SW, Lim KH, Kim KS, Kim SO, Hyeon T. Development of a T1 2. Hawrylak N, Ghosh P, Broadus J, Schlueter C, Greenough WT, Lauter- contrast agent for magnetic resonance imaging using MnO nanopar- bur PC. Nuclear magnetic resonance (NMR) imaging of iron oxide- ticles. Angew Chem Int Ed Engl 2007;46:5397–5401. labeled neural transplants. Exp Neurol 1993;121:181. 22. Walczak P, Kedziorek DA, Gilad AA, Barnett BP, Bulte JW. Applica- 3. Bulte JW, Hoekstra Y, Kamman RL, Magin RL, Webb AG, Briggs RW, Go bility and limitations of MR tracking of neural stem cells with asym- KG, Hulstaert CE, Miltenyi S, The TH, Deleij L. Specific MR imaging of metric cell division and rapid turnover: the case of the Shiverer dys- human lymphocytes by monoclonal antibody-guided dextran-magne- myelinated mouse brain. Magn Reson Med 2007;58:261–269. tite particles. Magn Reson Med 1992;25:148–157. 23. Gilad AA, McMahon MT, Walczak P, Winnard PT Jr, Raman V, van 4. Bulte JW, Kraitchman DL. Iron oxide MR contrast agents for molecular Laarhoven HW, Skoglund CM, Bulte JW, van Zijl PC. Artificial reporter and cellular imaging. NMR Biomed 2004;17:484–499. gene providing MRI contrast based on proton exchange. Nat Biotechnol 5. Bulte JW, Douglas T, Witwer B, Zhang SC, Strable E, Lewis BK, Zy- 2007;25:217–219. wicke H, Miller B, van Gelderen P, Moskowitz BM, Duncan ID, Frank 24. Gilad AA, Winnard PT Jr, van Zijl PC, Bulte JW. Developing MR JA. Magnetodendrimers allow endosomal magnetic labeling and in vivo reporter genes: promises and pitfalls. NMR Biomed 2007;20:275– tracking of stem cells. Nat Biotechnol 2001;19:1141–1147. 290.