HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 2006-0722v1 25/8/2128    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rosen, A. B. Articles by Cohen, I. S. Search for Related Content PubMed PubMed Citation Articles by Rosen, A. B. Articles by Cohen, I. S.

TRANSLATIONAL AND CLINICAL RESEARCH: MESENCHYMAL STEM CELLS SERIES

Finding Fluorescent Needles in the Cardiac Haystack: Tracking Human Mesenchymal Stem Cells Labeled with Quantum Dots for Quantitative In Vivo Three-Dimensional Fluorescence Analysis

^aDepartment of Biomedical Engineering, State University of New York at Stony Brook, Stony Brook, New York, USA;
^bDepartment of Physiology and Biophysics, State University of New York at Stony Brook, Stony Brook, New York, USA;
^cDepartment of Pharmacology, Columbia University, New York, New York, USA;
^dDepartment of Biomedical Engineering, Worcester Polytechnic Institute, Worcester, Massachusetts, USA;
^eInstitute for Molecular Cardiology, State University of New York at Stony Brook, Stony Brook, New York, USA;
^fCenter for Molecular Therapeutics, Columbia University, New York, New York, USA

Key Words. Human mesenchymal stem cells • Cell labeling • Cell tracking • Nanotechnology • Imaging

Correspondence: Amy B. Rosen, Ph.D., Institute for Molecular Cardiology, State University of New York at Stony Brook, Stony Brook, New York 11794, USA. Telephone: 631-444-7391; Fax: 631-444-3432; e-mail: amybrosen{at}gmail.com

Received November 8, 2006; accepted for publication May 4, 2007.
First published online in STEM CELLS EXPRESS   May 10, 2007.

	ABSTRACT

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
Stem cells show promise for repair of damaged cardiac tissue. Little is known with certainty, however, about the distribution of these cells once introduced in vivo. Previous attempts at tracking delivered stem cells have been hampered by the autofluorescence of host tissue and limitations of existing labeling techniques. We have developed a novel loading approach to stably label human mesenchymal stem cells with quantum dot (QD) nanoparticles. We report the optimization and validation of this long-term tracking technique and highlight several important biological applications by delivering labeled cells to the mammalian heart. The bright QD crystals illuminate exogenous stem cells in histologic sections for at least 8 weeks following delivery and permit, for the first time, the complete three-dimensional reconstruction of the locations of all stem cells following injection into the heart.

Disclosure of potential conflicts of interest is found at the end of this article.

	INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
The past decade has seen rapid advances in the use of embryonic and adult stem cells for tissue regeneration and repair in the heart [1–3]. These cells may have the potential to differentiate into mature cardiac cells or promote native repair through angiogenesis, recruitment of host stem cells, or induction of myocytes into the cell cycle [2, 4–6]. However, supporting studies are not without controversy [7, 8]; most have been unable to adequately track delivered stem cells with sufficient resolution in large animals. The ability to account for exogenous stem cells after delivery to animal models is important not only for determining the overall efficacy of intended treatments but also to rule out potentially dangerous side effects. For example, embryonic [9] and adult [10] stem cells have been injected into the myocardium as biological pacemakers; if these cells stray from the delivery site, they could generate focal arrhythmias with potentially fatal effects. Embryonic stem cells also have documented tumorigenic potential [11]. Thus, it becomes increasingly important to develop optimal tracking methods to identify delivered cells in vivo.

Traditional tracking agents such as green fluorescent protein (GFP) or fluorescent dyes fail to illuminate delivered cells above high levels of autofluorescence in the heart [12]. Secondary staining as used to detect LacZ or amplify GFP generates false positives and would also involve painstaking efforts to identify the exogenous cells in hundreds of tissue sections. More recently, cells have been labeled with inorganic particles for detection by magnetic resonance imaging (MRI) or positron emission tomography (PET) [13], but these imaging approaches can resolve no fewer than thousands of cells.

None of the existing tracking techniques offers the ability to unambiguously identify delivered cells in vivo with single-cell resolution using relatively high-throughput approaches (i.e., no secondary staining). Here, we report a novel approach to track human mesenchymal stem cells (hMSCs) using intracellular quantum dots (QDs). QDs are highly fluorescent semiconductor nanoparticles that possess unique optical properties [14, 15]. We demonstrate that single QD-hMSCs can be easily identified in histologic sections to determine their location for at least 8 weeks following delivery in vivo. Furthermore, we have utilized this approach to present for the first time a complete three-dimensional reconstruction of an in vivo stem cell "node."

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
Cell Culture
hMSCs were obtained from Clonetics/BioWhittaker (Walkersville, MD), and passages p3–p7 were used for all in vitro and in vivo experiments. Cells were grown on polystyrene tissue culture dishes and maintained at 37°C in humidified 5% CO₂ in mesenchymal stem cell growth medium (MSCGM) supplemented with L-glutamine, penicillin, and serum (MSCGM BulletKit; Cambrex, Walkersville, MD, http://www.cambrex.com). Cells were replated for passaging once every 2 weeks. For isolation of canine cardiac myocytes, adult mongrel dogs were intravenously injected with 80 mg/kg body weight sodium pentobarbital according to an approved protocol. Hearts were then removed and placed in a cold, high-potassium Tyrode solution [16]. Myocytes were isolated using a modified Langendorff system with digestion via Worthington type II collagenase [17], cultured onto laminin-coated glass coverslips, and maintained in Dulbecco's modified Eagle medium (DMEM) with 1% penicillin/streptomycin. Human umbilical vein endothelial cells (hUVECs) were obtained from Cambrex and cultured in their endothelial growth medium.

Quantum Dot Loading
Three approaches were used for loading hMSCs with QDs. First, a nucleofection protocol was followed to electroporate approximately 5 x 10⁵ hMSCs in 8.2 nM QD (Qdot 655 ITK carboxyl quantum dots; Invitrogen, Carlsbad, CA, http://www.invitrogen.com; catalog number Q21321MP) solution (supplemented Human MSC Nucleofector Solution; amaxa Inc., Gaithersburg, MD, http://www.amaxa.com; catalog number VPE-1001). After electroporation, cells were replated in complete MSCGM onto polystyrene tissue culture dishes. Second, a commercially available kit was used to load the cells with QDs via a carrier protein (Qtracker 655 Cell Labeling Kit; Invitrogen; catalog number Q25021MP). Briefly, 10 nM labeling solution was prepared according to kit directions, and approximately 0.2 ml was added to a 100-mm polystyrene tissue culture dish containing roughly 5 x 10⁵ cells. The cells were incubated at 37°C for 45–60 minutes, after which time they were washed twice with complete MSCGM. The third (and optimal) loading technique will be referred to as passive loading. Cells were grown to 85% confluence on polystyrene tissue culture dishes. An 8.2-nM solution of 655 ITK carboxyl QDs was prepared in complete MSCGM and vortexed for 60 seconds. Cells were washed once in phosphate-buffered saline (PBS) and incubated in the QD solution for up to 24 hours at 37°C.

Quantum Dot Validation and Mechanistic Experiments
After the incubation period, cells were washed twice in PBS, and fresh MSCGM was replaced. Loading efficiency was analyzed visually from a set of images of QD-hMSCs as well as by flow cytometry. In the first method, 181 cells from four QD fluorescence and phase contrast overlay images were studied and identified as either QD-positive or -negative. Loading determination via flow cytometry was assessed using the following protocol: hMSCs were loaded with QDs for 24 hours as described above. After the loading period, cells were washed twice in PBS, trypsinized, and resuspended in PBS with 5% fetal bovine serum. Cells were then stained with 7-amino-actinomycin D (to determine viability) and subsequently analyzed using a LSR II true multiparameter flow cytometer analyzer (with custom 655-nm filter; BD Biosciences, San Diego, http://www.bdbiosciences.com). Four sets of QD-hMSCs (and unloaded hMSCs for control), each containing a minimum of 17,000 cells, were analyzed. The intensity range for control cells was set such as to include at least 98% of the viable cells. The same technique was used to scan the QD-hMSCs, and determinations of QD-positive status were based on viable cells in the intensity range above that set for control.

A number of additional in vitro experiments were performed. To determine the degree of loading after repeated cell divisions, cells were passaged three times at 1:4 for a minimum of five divisions over 44 days. For one set of experiments intended to determine the mechanism of loading, hMSCs were passively exposed to QD incubation medium for 7 hours at either 4°C or 37°C. In another approach, cells were passively exposed to QD incubation medium for 12 hours either in MSCGM or 125 µM colchicine (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com; product number C9754) in MSCGM. To determine whether canine cardiac myocytes would take up QDs, cultured myocytes were incubated for up to 24 hours in DMEM to which the lysate from approximately 10⁴ QD-hMSCs was added.

Proliferation Assay
A population of hMSCs was evenly split for passaging, and both dishes were grown to 85% confluence. One dish was passively loaded with QDs as described above, whereas the other received a medium exchange. The following day (after 24 hours of loading), both dishes were washed and replated at equal concentrations into 12 wells each of a 96-well dish, and cells were allowed to grow for 3 days. The mitochondrial dehydrogenase assay (KK Biomed, Salt Lake City, http://www.kkbiomed.com) was carried out according to instructions provided by the company. The absorbance of each of the samples was measured at 595 nm using a POLARstar OPTIMA microplate reader (BMG Labtech, Durham, NC, http://www.bmglabtech.com).

Differentiation Experiments
Induction was performed using adipogenic and osteogenic kits available through Cambrex (Adipogenic Differentiation Medium, PT-3004; Osteogenic Differentiation Medium, PT-3002). All experiments were performed in triplicate on both QD-hMSCs and hMSCs.

For adipogenesis, labeled and unlabeled cells were plated at 2 x 10⁴ cells per cm² tissue culture surface area and fed every 2–3 days with MSCGM until cultures reached 100% confluence (5–13 days). Cells were fed on the following regimen for a total of three cycles: 3 days with supplemented Adipogenic Induction Medium followed by 1–3 days with Adipogenic Maintenance Medium. Control hMSCs were fed with Adipogenic Maintenance Medium at all times. After the three cycles, all cells were cultured for another week in Adipogenic Maintenance Medium. Cells were analyzed using light microscopy for characteristic lipid vacuole formation. MATLAB (MathWorks, Natick, MA, http://www.mathworks.com) algorithms were designed to determine percentage of images occupied by adipocytes.

For osteogenesis, cells were plated at 3 x 10³ cells per cm² tissue culture surface area and cultured overnight in MSCGM. Cells were then fed with Osteogenesis Induction Medium with replacement medium every 3–4 days for 2–3 weeks. Noninduced control cells were fed with MSCGM on the same schedule. Cells were analyzed using light microscopy for characteristic cobblestone appearance.

Gene Transfections
For some experiments, hMSCs were transfected with pIRES-EGFP (4 µg, Fig. 3), HCN2-pIRES-EGFP (4 µg, Fig. 3), or Wnt5A (4 µg, pUSEamp; Upstate, Charlottesville, VA, http://www.upstate.com; Fig. 4) plasmids using the amaxa biosystems nucleofection technique [10].

Patch Clamping
Whole cell patch clamping was executed as previously described [10]. Patch electrode resistance was 4–6 M. The pipette solution was filled with (in mM) K-aspartate 120, Mg-ATP 3, EGTA 10, and HEPES 5 (pH adjusted to 7.2 with KOH). The external solution contained (in mM) NaCl 137.7, KCl 5.4, NaOH 2.3, CaCl₂ 1.8, MgCl₂ 1, Glucose 10, and HEPES 5 (pH adjusted to 7.4 with NaOH). Recordings were made at room temperature.

Visualization of QD-hMSCs
In vitro experiments were performed in polystyrene tissue culture dishes. For typical visualization, cells were replated onto CC2-coated glass chamber slides (Lab-Tek; Nunc, Rochester, NY, http://www.nuncbrand.com). Several hours after replating, slides were rinsed in PBS and then fixed in 4% paraformaldehyde (PFA) for 15 minutes. Slides were rinsed again in PBS for 5 minutes and then incubated in 1 µM Hoechst 33342 nuclear dye (Cambrex) for 20 minutes. They were then washed in PBS for 5 minutes and placed in dH₂O for 20–30 seconds. The slides were rinsed successively in 30%, 70%, 95%, and 100% ethanol, each for 30 seconds, and then placed in 100% toluene for 30 seconds. Finally, slides were mounted in CytoSeal 60 (Electron Microscopy Sciences, Hatfield, PA, http://www.emsdiasum.com/microscopy) containing 1% trioctylphosphine (Sigma), with coverslips allowed to set overnight. Images were acquired on an inverted Zeiss Axiovert deconvolution microscope with AxioCam MRm CCD camera using a filter customized for 655-nm QD emission (Omega Optics Inc., Austin, TX, http://www.omegaoptics.com; XF3305; excitation at approximately 420 nm); the 4,6-diamidino-2-phenylindole (DAPI) filter set was used to visualize Hoechst 33342-stained nuclei. For some images, z-stacks were obtained at multiple focal planes and subsequently deconvolved using AxioVision (version 4.3; Carl Zeiss, Jena, Germany, http://www.zeiss.com). These stacks were then reassembled (in ImageJ, version 1.32j, NIH) into single two-dimensional images based on fluorescent pixels with maximum intensity in each section. All additional image processing was carried out using custom MATLAB algorithms (MATLAB 6.5 and 7.0) or in ImageJ. For some experiments, imaging was performed on live cells using an Olympus inverted fluorescence microscope (Olympus IX51, DP70 camera; Olympus, Tokyo, http://www.olympus-global.com) with GFP and Texas Red (for QD imaging) filter sets.

In Vivo
All animals received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication number 85-23, revised 1985). The animal protocols used were approved by the Institutional Animal Care and Use Committee at SUNY Stony Brook.

Patch Preparation
hMSCs were loaded with QDs as described above and subsequently transfected with the Wnt5A plasmid. A 15 x 30 x 0.1 mm acellular extracellular matrix (ECM) patch (porcine urinary bladder matrix; ACell, Jessup, MD, http://www.acell.com) was rinsed two times in PBS for 10 minutes each. The patch was then soaked in MSCGM for 15 minutes, after which time the medium was removed. QD-Wnt5A-hMSCs were trypsinized, resuspended in MSCGM, and seeded directly onto the ECM. The patch was returned to 37°C for approximately 12 hours prior to implantation.

Canine Patch Implants
Patches were implanted as described previously [18]. Briefly, a thoracotomy was used to expose the heart. A vascular clamp was then used to isolate a region of the right ventricular free wall. A full thickness defect was surgically induced and an hMSC-seeded scaffold was used to replace it. The chest was closed and the animal was allowed to recover. Animals were sustained under veterinary care and humanely terminated by an approved protocol at 8 weeks with pentobarbital.

Rat Heart Injections
QD-hMSCs were prepared as described above. Twenty-four hours after QD incubation, cells were washed twice in PBS, trypsinized, and resuspended for a final cell concentration of approximately 10⁵ cells per 10 µl in DMEM at 4°C. The cell solution was stored on ice until injection. Rats (5 months old; Charles River Laboratories, Wilmington, MA, http://www.criver.com) were anesthetized with ketamine/xylazine intraperitoneally, intubated, and maintained on inhaled isoflurane (1.5%–2%) for the duration of the experiment. A left thoracotomy was performed at the fourth or fifth intercostal space. A 5–0 prolene suture was used to place a superficial stitch in the epicardium as a fiducial marker. Ten µl of cell solution or cell lysate was injected into the free left ventricular wall apical to the suture, and then a small drop of surgical grade tissue adhesive (Nexaband; Webster Veterinary, Sterling, MA, https://www.jawebster.com) was applied over the injection site. The thorax was closed, and rats were returned to their cages for either 1 hour or 1 day for whole cell injections or either 1 hour or 1 week for the lysed QD-hMSC injections. Euthanasia was performed either in a CO₂ chamber or by administering pentobarbital (100 mg/kg body weight injected intraperitoneally) and subsequent cardiectomy.

Preparation of Tissue Samples
Immediately after explantation, tissue samples were rinsed in isotonic saline and then fixed in 4% PFA for 24 hours. After fixation, tissue was cryopreserved in an isotonic 30% sucrose solution for at least 24 hours. Gross photographs were obtained of tissue samples with sutures in situ to identify the cell delivery zone (either patch borders or injection site). After suture removal, tissue was embedded in freezing matrix (Jung tissue embedding matrix; Leica, Heerbrugg, Switzerland, http://www.leica.com) and stored at –20°C. Ten-µm tissue sections were cut on a cryotome, transferred to Suprafrost glass slides, and stored at –20°C. Slides were either imaged without mounting under glass or stained with Hoechst 33342 dye and mounted as described above.

Immunohistochemistry
Staining for CD31 was performed as follows: 10-µm histologic sections or CC2-coated chamber slides containing cells (hUVECs or hMSCs) were hydrated with PBS and permeabilized with 0.5% Tween 20 in 1x Tris-buffered saline (TBS) followed by 0.25% Triton X-100 in TBS. Samples were blocked with normal horse serum (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) and then incubated for 4 hours in fluorescein isothiocyanate-conjugated anti-human CD31 (Diaclone, Stamford, CT, http://www.diaclone.com). Control sections were incubated in PBS instead of the antibody. Sections were washed in PBS, incubated with 1 µM Hoechst 33342 for 20 minutes (Invitrogen), and then washed in PBS before mounting with Vectashield (Vector).

Statistics
All data are listed as mean ± standard deviation. Data sets were compared by a Student's t test with p < .05 considered significant.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
Loading of hMSCs Is Optimized by Passive Incubation with Negatively Charged QDs and Is Blocked by Inhibitors of Endocytosis
Optimal use of QDs for tracking hMSCs requires nearly 100% cell survival after loading and that loaded cells behave similarly to unloaded cells. Potential approaches to loading populations of cells include electroporation [19], lipid vehicles [19–21], and passive (receptor-mediated or unmediated) incubation [22–27]. We examined loading by some of these methods using QDs with either positively or negatively charged surface conjugations. Electroporation was least effective, loading only a small fraction of the hMSC population and causing appreciable cell death (data not shown). Receptor-mediated uptake was more effective (Fig. 1A) but still nonuniform. Unfortunately, both methods resulted in marked aggregation of the QDs in the perinuclear spaces of the hMSCs over time. Passive incubation with naked QDs (655-nm peak emission wavelength) having carboxylic acid derivatizations on their polymer surfaces (net negatively charged) was most effective (Fig. 1B, 1C). The pattern of loading was diffusely cytoplasmic (Fig. 1D). Virtually all hMSCs were loaded (>98% of 181 cells analyzed visually in four images and >96% of >17,000 cells per set analyzed by flow cytometry, n = 4 sets, Fig. 1E). Interestingly, when incubations were attempted using smaller-core-sized QDs (525-nm peak emission wavelength) with either positively or negatively charged surface conjugations, we were unable to achieve similar levels of loading (data not shown). We investigated the mechanism of QD loading by employing two protocols that reduce endocytosis: exposure to low temperature for 7 hours (4°C, data not shown) and application of colchicine (125 µM, an inhibitor of microtubule aggregation [28]) for 24 hours prior to and during incubation with QDs (Fig. 1F). In each case, there was a dramatic reduction in QD uptake. Finally, we investigated the time course of QD loading by passively incubating cells with QDs for 1, 3, 7, 12, and 24 hours. Intracellular QDs were barely detectable at 1 hour of incubation (Fig. 1G), easily detectable after 3 hours (Fig. 1H), and quite bright at 24 hours (Fig. 1I), prompting us to choose an incubation range of 12–24 hours for most experiments. A summary of conditions used to optimize loading is shown in Table 1.

View larger version (31K): [in this window] [in a new window]   Figure 1. Quantum dots (QDs) can be loaded uniformly into hMSCs by passive incubation. QD loading was achieved by receptor-mediated uptake or passive incubation with naked dots. Panels (A) and (B) show images of QD fluorescence (655 nm, red) with phase contrast overlays. (A): Use of the receptor-mediated-based Qtracker kit resulted in nonuniform cellular loading with perinuclear aggregation. (B): In contrast, passively incubating hMSCs in naked QD medium results in nearly 100% loading with a pattern that extends to the cell borders. (C): The field in (B) is imaged for QD fluorescence without the phase overlay to demonstrate homogeneity and brightness. The intracellular QD cluster distribution is diffusely cytoplasmic (C, D) and largely excludes the nucleus (blue, Hoechst 33342 dye). (E): QD loading efficiency was analyzed using flow cytometry. The threshold for plain hMSCs (gray line) was set such that the intensity range encompassed at least 98% of the control cells (red arrow indicates upper bound of control range). QD-positive status was designated for all cells in the QD-hMSC sample having intensities above the range set for the control group. In four experiments, QD-positive cells (black line) were found in 96% of over 17,000 viable cells. (F): When colchicine-conditioned cells are incubated with colchicine-containing QD medium, the uptake is dramatically reduced. The cells in panel ([F], colchicine) and ([F], inset, no colchicine) can be directly compared, as the incubation periods were identical and the cells were imaged using the same microscope and camera settings. (G–I): hMSCs continually take up QDs from the incubation medium. Cells incubated in QD medium for (G) 1, (H) 3, and (I) 24 hours are all imaged using the same microscope and camera settings. Images are grayscale for clarity. Because of the different levels of loading, the exposure time (380 milliseconds) used to image all three samples is clearly too high for the 24-hour-incubated cells; most of the QD clusters in the image are overexposed. Scale bar (A, B) = 50 µm. Scale bar (C) = 100 µm. Scale bar (D, F–I) = 25 µm. Abbreviations: hMSCs, human MSCs; QD-hMSCs, quantum dot-human MSCs.

View larger version (22K): [in this window] [in a new window]   Figure 2. Quantum dots (QDs) retain their brightness and cytoplasmic distribution for up to 6 weeks in vitro and are not transferred to unloaded cells. QD-hMSCs were fixed onto slides and stained with Hoechst 33342 dye as described in Materials and Methods. The cells were imaged for QD fluorescence at (A) 2 days, (B) 16 days, and (C) 44 days after loading. Only the Hoechst (blue) channels of these images have been postprocessed to enhance contrast; QD channels (red) are displayed as imaged. As cells divide, they split their cytoplasmic contents to each daughter cell, diluting the ultimate concentration of QDs in progeny over time. Therefore, exposure times of excitation light need to be increased to optimally capture QD images of cells that have been through multiple divisions. Microscope and camera settings are the same for images (A) and (B), but in (C), cells are imaged for 655-nm (red) emission using triple the exposure time. The QDs in (A) are overexposed at this setting, with some dots clearly saturating the image. (D): QD-loaded and plain hMSCs proliferate similarly, as measured by a mitochondrial dehydrogenase assay (n = 12). (E): Green (green fluorescent protein [GFP] transfected) and red (QD-loaded) hMSCs in direct apposition were imaged live to look for QD transfer to neighboring cells (colors have been enhanced for contrast). In a separate experiment, QD-hMSCs were mechanically lysed and then added to cultured canine myocytes for 24 hours. (F): A myocyte (right) sits near a live QD-hMSC that survived the lysis and attached to the coverslip. Floating QD clusters from the lysed cells are apparent in the medium but were not internalized by the myocytes. Live cell images in (E) and (F) were acquired on an Olympus inverted fluorescence microscope using a GFP and a Texas Red filter, which does not optimally image the 655-nm QD signal. Scale bar (A–C) = 25 µm. Scale bar (E, F) = 50 µm. Abbreviations: a.u., arbitrary units; hMSCs, human MSCs; QD-hMSCs, quantum dot-human MSCs.

QDs Are Not Taken Up by Adult Cardiac Myocytes in Culture
hMSCs have been shown to enhance cardiac regeneration in animal trials [4]. If QDs are used to track the fate of stem cells delivered to the heart, myocytes must not take the dots up from the extracellular space should these exogenous cells die in their vicinity. To simulate the in vivo situation of dying hMSCs, we exposed cultured cardiac myocytes to the cell lysate from mechanically disrupted QD-hMSCs for 24 hours. Figure 2F provides one example, demonstrating that the myocytes did not take up QDs. We also performed an equivalent control using lysed cells in vivo, which is discussed below.

QD-Loaded hMSCs Can Be Transfected to Overexpress Genes
Because hMSCs are an attractive vehicle for gene delivery to the heart [10], we queried whether the presence of intracellular QDs would affect expression of exogenous genes. QD-hMSCs were transfected with the HCN2-pIRES-EGFP plasmid. The HCN2 gene expresses a time-dependent inward current, which is the basis of cardiac pacemaker activity. We have used this plasmid previously with hMSCs as the cellular vehicle to create a biological pacemaker in the canine heart [10]. After 48 hours, QD-loaded cells were visualized for GFP expression and compared with control hMSCs that underwent the same transfection protocol but were not first exposed to QDs. GFP-positive cells from each group were then selected for patch clamping to record membrane currents (Fig. 3A, 3B). QD-hMSCs expressed the HCN2 gene and generated a family of pacemaker currents similar to those recorded in unloaded cells. The current amplitudes recorded at –150 mV for both control and QD-hMSCs were –1459.48 ± 616.83 and –1352.68 ± 864.70, respectively (p > .05, n = 5 per group).

View larger version (25K): [in this window] [in a new window]   Figure 3. The presence of intracellular quantum dots (QDs) does not affect ability of cells to overexpress genes after transfection. QD-human MSCs (hMSCs) and plain hMSCs were each transfected with the HCN2-pIRES-EGFP plasmid via electroporation. Two days after transfection, both groups of cells expressed similar levels of green fluorescent protein (GFP) (data not shown), and cells expressing GFP from both groups were patch clamped to record the HCN2-induced currents. The currents provided were from a holding potential of –40 mV and included steps between –40 mV and –160 mV in –10 mV increments. Similar levels of HCN2-induced current were recorded from (A) unloaded and (B) QD-loaded hMSCs. Additionally, the electroporation process did not alter the cytoplasmic distribution of QDs ([B], inset). Imaging for these experiments was carried out on the Olympus microscope using the Texas Red filter to visualize QDs in live cells; this filter set does not optimally visualize QD loading. Scale bar on (B) inset = 50 µm. Abbreviations: nA, nanoampere; S, second.

View larger version (63K): [in this window] [in a new window]   Figure 4. Quantum dots (QDs) do not interfere with differentiation capacity of human MSCs (hMSCs) in vitro. QD-hMSCs or unloaded hMSCs were induced to differentiate in vitro along adipogenic and osteogenic lineages. After the adipogenic induction period, both (A) unloaded hMSCs and (B) QD-hMSCs displayed characteristic adipocyte morphology with prominent lipid vacuoles. The percentage of differentiated versus undifferentiated cells was similar between these two groups (44.9% in [A]; 40.4% in [B]). (C): At high power, adipocytes from the QD-hMSC group are seen with QDs (red fluorescence) interspersed between lipid vacuoles. After the osteogenic induction period, both (D) unloaded hMSCs and (E) QD-hMSCs that were initially spindle-shaped adapted a more cobblestone-like morphology typical of osteocytes and tended to cluster on the dish. (F): At high power, a cobblestone-shaped osteocyte from the QD-hMSC group shows retention of the QD label (red) after the differentiation process. Scale bars on (A, B, D, E) = 200 µm. Scale bars on (C, F) = 50 µm.

View larger version (30K): [in this window] [in a new window]   Figure 5. Quantum dot-human MSCs (QD-hMSCs) can be delivered to the canine heart on an extracellular matrix (ECM) scaffold and identified 8 weeks later; QDs do not interfere with differentiation in vivo. QD-hMSCs were delivered to the canine ventricle via implantation of an ECM patch. Eight weeks later, tissue was explanted and fixed. Panel (A) shows fixed tissue from one animal with a blue line circumscribing the region analyzed (and imaged transmurally in panel [C]) and a black dotted ellipse approximating the patch borders. Straight dark black lines in the image are dissecting pins that were used to secure the tissue while photographing. The region outlined in blue in (A) was frozen and sectioned transmurally at 10 µm and (B) imaged for QD fluorescence (655 nm) and phase contrast. (C): The plane of section is transmural from epicardium (top) to endocardium (bottom); a green circle highlights the region where QDs were found (imaged in [B]). Some tissue sections were stained with Hoechst 33342 to visualize nuclei. Panel (B), inset, shows QD-hMSCs amid endogenous tissue (asterisks denote endogenous nuclei). (D): After 8 weeks, some of these QD-positive cells (red) express the endothelial marker platelet/endothelial cell adhesion molecule (PECAM)-1 (green), suggesting differentiation of these cells along an endothelial lineage. A high-power view of QD-positive cells with colocalized PECAM-1 expression is shown in (D), inset. In these images, exposure times for obtaining QD fluorescence were selected to avoid overexposure in the regions of higher QD density and may therefore fail to visualize all QD-positive cells. Controls include: (E): Host canine endothelial cells seen lining the lumen of a blood vessel do not stain positively for the anti-human CD31, emphasizing the species-specificity of the antibody (same microscope and camera settings for imaging as in [D]). (F): Cultured human endothelial cells express but (G) cultured hMSCs do not express the anti-human CD31. (H): Negative control human endothelial cells without exposure to the fluorescein isothiocyanate-conjugated anti-CD31 antibody. Insets in (F), (G), and (H) show higher power views (630x) of individual cells. Images in (F), (G), and (H) and respective insets acquired using same microscope and camera settings per objective. Scale bar on (A, C) = 20 mm. Scale bar on (B) = 50 µm, inset = 10 µm. Scale bar on (D) = 20 µm, inset = 5 µm. Scale bar on (E) = 20 µm. Scale bar on (F, G, H) = 50 µm, insets on (F, G, H) = 20 µm.

QDs Are Not Internalized by Cardiac Cells In Vivo
To further ensure that we would not be misled by false positives, we performed a set of experiments to determine whether native myocardial cells internalize QDs in vivo. QDs can exist extracellularly if QD-hMSCs die and leak their contents. Therefore, we injected into the rat ventricle a suspension of approximately 100,000 QD-hMSCs that were mechanically disrupted to cause cell lysis and terminated the animals at either 1 hour (n = 2) or 1 week (n = 2). We did not observe QDs in any cell type in these hearts. This finding is expected, as free carboxylated QDs will be removed by the reticuloendothelial system in less than 1 hour [31].

The Number and Distribution of QD-hMSCs Injected into a Rat Heart Can Be Reconstructed in Three Dimensions
We previously injected 1 million hMSCs into canine myocardium to create a biological pacemaker and used traditional means of identifying these cells in histologic sections (GFP and secondary staining) [10]. Although we were able to demonstrate the presence of some of our delivered cells, it was not possible to reconstruct their three-dimensional locations. Such added information could help us understand how these HCN2-transfected stem cells generated pacemaker activity and assess their potential to develop unwanted arrhythmic events. With this need in mind, we performed a series of experiments in rats to enumerate delivered cells in vivo and reconstruct their spatial distribution in three dimensions (3D). Approximately 100,000 QD-hMSCs were injected into the left ventricular free wall. Hearts were harvested at either 1 hour or 1 day after injection. Serial 10-µm transverse sections were imaged for QD fluorescence (Fig. 6A, 6E). Using algorithms described in the supplemental online Materials and Methods, these fluorescence images were filtered and thresholded to generate binary maps of QD-positive zones from all of the tissue sections (Fig. 6B). The spatial locations of QD-hMSCs were identified from the series of binary maps and visualized in 3D (Fig. 6C, 6D and supplemental online movies 1 and 2). Our algorithms permit enumeration of the total number of QD-hMSCs in the whole heart from these models (approximately 50,000 at 1 hour and 30,000 at 1 day). We also computed a distance parameter to characterize the distribution of cells based on the distance between individual cells and the centroid of the total stem cell mass. Most of the cells were clustered in close proximity (85% of cells within 1.5 mm at 1 hour and 95% within 1.5 mm at 24 hours, Fig. 6F).

View larger version (51K): [in this window] [in a new window]   Figure 6. Quantum dots (QDs) can be used to identify single human MSCs (hMSCs) after injection into the rat heart and further used to reconstruct the three-dimensional (3D) distribution of all delivered cells. Rat hearts were injected with QD-hMSCs. Fixed, frozen sections were cut transversely (plane shown in [B], inset) at 10 µm and mounted onto glass slides. Sections were imaged for QD fluorescence emission (655 nm) with phase overlay to visualize tissue borders. QD-hMSCs can be visualized at (A) low power and (inset) high power (Hoechst 33342 dye used to stain nuclei blue). In (A), inset, endogenous nuclei can be seen adjacent to the delivered cells in the mid myocardium (arrows). Serial low power images were registered with respect to one another and (B) binary masks were generated where white pixels depict the QD-positive zones in the images. The vertical line in (B), inset, represents the z-axis, which has a zero value at the apex of the heart. The binary masks for all of the QD-positive sections of the heart were compiled and used to generate the 3D reconstruction of delivered cells in the tissue. QD-hMSCs remaining in the tissue adhesive on the epicardial surface (and not within the cardiac syncytium) were excluded from the reconstruction. (C): QD-hMSC reconstruction in an animal that was terminated 1 hour after injection. (D): Reconstruction from an animal euthanized 1 day after injection with orientation noted in inset. Our reconstructions in (C) and (D) do not account for all of the approximately 100,000 hMSCs delivered through the needle. Some of these cells undoubtedly leaked out of the needle track, whereas others may not have survived the injection protocol. The views of both reconstructions (C) and (D) are oriented for optimal static visualization (and also have different scales and are situated at different positions along the z-axis depending on the distance of the injection site from the apex of the heart); see supplemental online movies for complete 3D representations from all perspectives. (E): One day after injection into the heart, the pattern of QD-hMSCs is well organized and appears to mimic the endogenous myocardial orientation (dotted white line highlights myofibril alignment). Complete representations of the spatial localization of QD-hMSCs in the heart permit further quantitative analyses. (F): One parameter that can be computed is the distance of individual cells from the centroid of the total cell mass. The plots show the percentage of cells at a distance less than or equal to x for both the 1-hour and 1-day rats. At both time points, most of the cells are within 1.5 mm of the centroid. Scale bar on (A) = 500 µm, inset = 20 µm. Scale bar on (B), inset = 1 cm. Scale bar on (E) = 500 µm. Abbreviations: endo, endogenous; epi, epicardial.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

Existing techniques for tracking cells include (a) transfecting cells in vitro with either fluorescent (i.e., GFP) [10] or nonfluorescent (i.e., ß-galactosidase) proteins [34], (b) employing fluorescence in situ hybridization (FISH) to identify delivered cells by their unique chromosomal content (i.e., male cells delivered to a female host) [35], (c) using species-specific surface markers to identify delivered cells in a xenographic host [10], (d) using commercially available fluorescent dyes to label cell cytoplasm [36] or nuclei [37], or, more recently, (e) labeling cells with inorganic particles (i.e., Feridex) [4, 36, 38] or radiotracers [13, 36, 39, 40]. The above can be grouped into two categories: those techniques requiring secondary staining to visualize cells in histologic sections (a–c) and those not requiring secondary staining (d, e). Although all of these approaches have advantages, they have generally focused on identifying the location and fate of a small minority of the delivered cells or produced lower resolution maps of the in vivo stem cell distribution.

Within the first category of techniques, transfection is inefficient and cannot label 100% of the cells without selection for stable cell lines. Loading cells with exogenous DNA also requires use of a viral or Lipofectamine vehicle or electroporation; each of these processes causes some cell death. Immunostaining—employed when tracking cells with FISH or by surface markers—generates some false positives [12], and interpretation of staining outcomes can be difficult in the absence of strict controls. Stem cells are also dynamic and may change their surface marker expression over time if they differentiate in vivo. Furthermore, staining results and image features are typically heterogenous across a given tissue section and between sections, making it important to carefully sample the specimens. In general, immunostaining is a low-throughput process that is not practical to account for all delivered cells.

The second category of cell tracking approaches—those not requiring secondary staining—represents a potentially higher-throughput avenue to identifying delivered cells. Fluorescent dyes, however, fail to demonstrate any advantage over other techniques since they must be taken up via lipid vehicles, are only retained for short periods of time, and are prone to photobleaching when imaged [41]. As with GFP, their emission spectra also overlap with that of host tissue autofluorescence. In the cardiac milieu, for example, both normal and damaged myocardium emit autofluorescence that can easily be misinterpreted as exogenous fluorescent cells. This phenomenon occurs in almost the entire visible spectral range [12, 42]. Some groups have labeled cells with the nuclear dye DAPI [36, 37], but this agent intercalates into DNA and can therefore interfere with normal cell function [41]. Recently, others have loaded stem cells with magnetic nanoparticles (superparamagnetic iron oxide [SPIO]) [36] or radiotracers [40] and imaged delivered cells in vivo with MRI or PET scanning. This offers the advantage of identifying clusters of labeled cells noninvasively but does not afford single-cell resolution. SPIO-loaded cells can be identified in histologic sections only by secondary staining techniques (i.e., Prussian blue reaction).

We have shown that quantum dots can be used as a reliable long-term tracking agent for identifying exogenous hMSCs in histologic sections. QDs have been used in a number of biological applications [14, 15, 21, 22, 25, 43, 44]. Among the most relevant advantages of labeling cells with QDs are their extreme brightness, photostability, and large Stokes shift. The latter feature means that a sample can be excited at any wavelength lower than its emission wavelength. One can take advantage of this feature by exciting in the low UV range (400 nm) while collecting a very tight spectrum of emitted light in the red range (i.e., 640–660 nm). This makes the collected signal QD-specific and is one of the ways in which imaging QDs in histologic sections mitigates the effects of autofluorescence.

When others have attempted to label populations of cells with QDs using receptor-mediated uptake, lipofection, or electroporation, the dots have aggregated around the nucleus [19, 26, 45]. Hsieh et al. recently correlated this behavior with significant impairment of cell function at the molecular level [46]. Our passive loading approach overcomes some of the potential problems seen with other QD loading techniques, resulting in uniform diffuse cytoplasmic labeling of populations of cells and avoiding perinuclear aggregation. This behavior is maintained for at least 6 weeks in vitro and 8 weeks in vivo. Effective loading depends on the size and charge of the QDs as well as the properties of the incubation medium. When loaded according to our protocol, QD-hMSCs have similar proliferative and differentiation capacities to unloaded hMSCs, suggesting that this labeling method is appropriate for fate-tracking studies in vitro and in vivo.

QDs can be easily imaged in the highly autofluorescent cardiac milieu because (a) custom filters selectively excite the sample at a wavelength far from the emission spectrum; (b) these same filters capture emitted light in a QD-specific range; (c) the bright signal requires very low exposure times to image the samples, making it less likely that autofluorescent emission will be collected; (d) QD emission in the red and near-infrared range scatters less in tissue samples; and (e) autofluorescence makes less of a contribution in the red and near-infrared range than at lower wavelengths of the spectrum [42]. The bright QD fluorescence and relatively high-throughput nature of our approach allow us to obtain hundreds of low-power images of QD-positive cells in sections of heart tissue. Because of the robustness of their fluorescence, QD-hMSCs can be reliably imaged from samples up to 8 weeks after in vivo delivery.

In order to locate our delivered cells, we sectioned several thousand microns worth of tissue and visually sampled slides at hundred-micron intervals. We identified the QD-positive regions several hundred microns into the tissue, emphasizing the importance of careful sampling; many existing studies demonstrate findings from a single histologic section but in the absence of an unambiguous marker, it is very unlikely that a random section will contain any delivered cells. Had secondary staining been necessitated on all of these sections—as would have been the case for traditional cell tracking approaches—the process of identifying the delivered cells would have taken many months as opposed to several days. LeGrice et al. have recently developed automated techniques for sectioning and subsequent confocal imaging of myocardial samples [47]. Such a system could be extended to image QD-hMSCs within a network of host tissue and even further streamline our approach.

After validating the use of QDs as a vital long-term stem cell tracking agent, we tested their potential to address important biological questions. Prior to this study, little had been documented about the spatial distribution of stem cells after injection into the heart. We exploited the relative ease of identifying QD-hMSCs in low-power images of histologic sections and designed custom computer algorithms to generate 3D reconstructions of the entire cell zone. Thus, we present the first complete single-cell resolution quantitative representations of the distribution of delivered stem cells in vivo. Furthermore, we induced a cardiac injury in canine ventricle via removal of a full thickness of myocardium and replaced this site with an ECM patch onto which QD-hMSCs were seeded. After 8 weeks, QD fluorescence was clearly visible in the injured tissue well above high levels of autofluorescence, and some of these cells had developed an endothelial phenotype. In another set of experiments not reported here, we loaded hMSCs with QDs and committed them to a cardiac lineage in vitro. We then delivered these cells to the canine on a patch using the ventricular defect model and demonstrated that the QD-containing cardiogenic cells differentiated into mature myocytes with normal sarcomere spacing [48, 49]. From these results, we conclude that QDs successfully illuminate delivered hMSCs in both normal and damaged myocardium and can be used to track stem cell location and fate after delivery.

The qualities of an ideal agent for tracking stem cells have been outlined by Frangioni and Hajjar [50] and include the criteria that the agent (a) be biocompatible, safe, and nontoxic; (b) not require any genetic modification or perturbation to the stem cell; (c) permit single cell detection at any anatomic location; (d) allow quantification of cell number; (e) have minimal or no dilution with cell division; (f) have minimal or no transfer to non-stem cells; (g) permit noninvasive imaging in the living subject over months to years; and (h) require no injectable contrast agent for visualization. We have demonstrated that QDs satisfy most of these criteria and represent a viable approach for long-term stem cell tracking.

	CONCLUSION

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
In summary, under the conditions described herein, populations of hMSCs can be uniformly and completely loaded with QD nanoparticles for long-term cell tracking in vivo. QD-hMSCs can be unambiguously imaged in histologic sections at least 8 weeks after delivery, and their spatial distribution can be reconstructed in 3D. A better understanding of the distribution and fate of delivered cells should facilitate assessment of safety and efficacy for any new stem cell therapy before it reaches human trials.

	DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
I. Cohen has acted as a consultant to and performed contract work for Guidant Corporation within the last two years. M. Rosen, R. Robinson, and P. Brink have performed contract work for Guidant Corporation within the last two years.

	ACKNOWLEDGMENTS

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 
We acknowledge Joan Zuckerman for invaluable technical assistance. This work has been partially supported by Boston Scientific Corporation. This work was supported by Grants HL67101, HL28958, DK60037, two SDGs from the AHA (one to S.V.D.; one to G.R.G.), NIH training Grant DK07521 (to D.J.K.), and a contract from Guidant Corporation.

	REFERENCES

Top Abstract Introduction Materials and Methods Results Discussion Conclusion Disclosure of Potential... Acknowledgments References

 

1. Klug MG, Soonpaa MH, Koh GY et al. Genetically selected cardiomyocytes from differentiating embronic stem cells form stable intracardiac grafts. J Clin Invest 1996;98:216–224.[Medline]

2. Yoon YS, Wecker A, Heyd L et al. Clonally expanded novel multipotent stem cells from human bone marrow regenerate myocardium after myocardial infarction. J Clin Invest 2005;115:326–338.[CrossRef][Medline]

3. Meyer GP, Wollert KC, Lotz J et al. Intracoronary bone marrow cell transfer after myocardial infarction: Eighteen months' follow-up data from the randomized, controlled BOOST (BOne marrOw transfer to enhance ST-elevation infarct regeneration) trial. Circulation 2006;113:1287–1294.[Abstract/Free Full Text]

4. Amado LC, Saliaris AP, Schuleri KH et al. Cardiac repair with intramyocardial injection of allogeneic mesenchymal stem cells after myocardial infarction. Proc Natl Acad Sci U S A 2005;102:11474–11479.[Abstract/Free Full Text]

5. Mangi AA, Noiseux N, Kong D et al. Mesenchymal stem cells modified with Akt prevent remodeling and restore performance of infarcted hearts. Nat Med 2003;9:1195–1201.[CrossRef][Medline]

6. Beltrami AP, Urbanek K, Kajstura J et al. Evidence that human cardiac myocytes divide after myocardial infarction. N Engl J Med 2001;344:1750–1757.[Abstract/Free Full Text]

7. Orlic D, Kajstura J, Chimenti S et al. Bone marrow cells regenerate infarcted myocardium. Nature 2001;410:701–705.[CrossRef][Medline]

8. Murry CE, Soonpaa MH, Reinecke H et al. Haematopoietic stem cells do not transdifferentiate into cardiac myocytes in myocardial infarcts. Nature 2004;428:664–668.[CrossRef][Medline]

9. Kehat I, Khimovich L, Caspi O et al. Electromechanical integration of cardiomyocytes derived from human embryonic stem cells. Nat Biotechnol 2004;22:1282–1289.[CrossRef][Medline]

10. Potapova I, Plotnikov A, Lu Z et al. Human mesenchymal stem cells as a gene delivery system to create cardiac pacemakers. Circ Res 2004;94:952–959.[Abstract/Free Full Text]

11. Thomson JA, Itskovitz-Eldor J, Shapiro SS et al. Embryonic stem cell lines derived from human blastocysts. Science 1998;282:1145–1147.[Abstract/Free Full Text]

12. Laflamme MA, Murry CE. Regenerating the heart. Nat Biotechnol 2005;23:845–856.[CrossRef][Medline]

13. Kraitchman DL, Tatsumi M, Gilson WD et al. Dynamic imaging of allogeneic mesenchymal stem cells trafficking to myocardial infarction. Circulation 2005;112:1451–1461.[Abstract/Free Full Text]

14. Bruchez M Jr, Moronne M, Gin P et al. Semiconductor nanocrystals as fluorescent biological labels. Science 1998;281:2013–2016.[Abstract/Free Full Text]

15. Chan WC, Nie S. Quantum dot bioconjugates for ultrasensitive nonisotopic detection. Science 1998;281:2016–2018.[Abstract/Free Full Text]

16. Isenberg G, Klockner U. Calcium tolerant ventricular myocytes prepared by preincubation in a "KB medium". Pflugers Arch 1982;395:6–18.[CrossRef][Medline]

17. Cordeiro JM, Greene L, Heilmann C et al. Transmural heterogeneity of calcium activity and mechanical function in the canine left ventricle. Am J Physiol Heart Circ Physiol 2004;286:H1471–H1479.[Abstract/Free Full Text]

18. Kochupura PV, Azeloglu EU, Kelly DJ et al. Tissue-engineered myocardial patch derived from extracellular matrix provides regional mechanical function. Circulation 2005;112 (suppl 9):I144–I149.[Medline]

19. Derfus AM, Chan WCW, Bhatia SN. Intracellular delivery of quantum dots for live cell labeling and organelle tracking. Adv Mater 2004;16:961–966.[CrossRef]

20. Dubertret B, Skourides P, Norris DJ et al. In vivo imaging of quantum dots encapsulated in phospholipid micelles. Science 2002;298:1759–1762.[Abstract/Free Full Text]

21. Voura EB, Jaiswal JK, Mattoussi H et al. Tracking metastatic tumor cell extravasation with quantum dot nanocrystals and fluorescence emission-scanning microscopy. Nat Med 2004;10:993–998.[CrossRef][Medline]

22. Gao X, Cui Y, Levenson RM et al. In vivo cancer targeting and imaging with semiconductor quantum dots. Nat Biotechnol 2004;22:969–976.[CrossRef][Medline]

23. Hoshino A, Hanaki K, Suzuki K et al. Applications of T-lymphoma labeled with fluorescent quantum dots to cell tracing markers in mouse body. Biochem Biophys Res Commun 2004;314:46–53.[CrossRef][Medline]

24. Zhang Y, Huang N. Intracellular uptake of CdSe-ZnS/polystyrene nanobeads. J Biomed Mater Res B Appl Biomater 2006;76:161–168.[Medline]

25. So MK, Xu C, Loening AM et al. Self-illuminating quantum dot conjugates for in vivo imaging. Nat Biotechnol 2006;24:339–343.[CrossRef][Medline]

26. Silver J, Ou W. Photoactivation of quantum dot fluorescence following endocytosis. Nano Lett 2005;5:1445–1449.[CrossRef][Medline]

27. Jaiswal JK, Mattoussi H, Mauro JM et al. Long-term multiple color imaging of live cells using quantum dot bioconjugates. Nat Biotechnol 2003;21:47–51.[CrossRef][Medline]

28. Piasek A, Thyberg J. Effects of colchicine on endocytosis and cellular inactivation of horseradish peroxidase in cultured chondrocytes. J Cell Biol 1979;81:426–437.[Abstract/Free Full Text]

29. Valiunas V, Doronin S, Valiuniene L et al. Human mesenchymal stem cells make cardiac connexins and form functional gap junctions. J Physiol 2004;555Pt 3:617–626.[Abstract/Free Full Text]

30. Miyahara Y, Nagaya N, Kataoka M et al. Monolayered mesenchymal stem cells repair scarred myocardium after myocardial infarction. Nat Med 2006;12:459–465.[CrossRef][Medline]

31. Ballou B, Lagerholm BC, Ernst LA et al. Noninvasive imaging of quantum dots in mice. Bioconjug Chem 2004;15:79–86.[CrossRef][Medline]

32. Britten MB, Abolmaali ND, Assmus B et al. Infarct remodeling after intracoronary progenitor cell treatment in patients with acute myocardial infarction (TOPCARE-AMI): Mechanistic insights from serial contrast-enhanced magnetic resonance imaging. Circulation 2003;108:2212–2218.[Abstract/Free Full Text]

33. Wollert KC, Meyer GP, Lotz J et al. Intracoronary autologous bone-marrow cell transfer after myocardial infarction: The BOOST randomised controlled clinical trial. Lancet 2004;364:141–148.[CrossRef][Medline]

34. Jackson KA, Majka SM, Wang H et al. Regeneration of ischemic cardiac muscle and vascular endothelium by adult stem cells. J Clin Invest 2001;107:1395–1402.[CrossRef][Medline]

35. Kajstura J, Rota M, Whang B et al. Bone marrow cells differentiate in cardiac cell lineages after infarction independently of cell fusion. Circ Res 2005;96:127–137.[Abstract/Free Full Text]

36. Kraitchman DL, Heldman AW, Atalar E et al. In vivo magnetic resonance imaging of mesenchymal stem cells in myocardial infarction. Circulation 2003;107:2290–2293.[Abstract/Free Full Text]

37. Thompson RB, Emani SM, Davis BH et al. Comparison of intracardiac cell transplantation: Autologous skeletal myoblasts versus bone marrow cells. Circulation 2003;108 (suppl 1):II264–II271.[Medline]

38. Dick AJ, Guttman MA, Raman VK et al. Magnetic resonance fluoroscopy allows targeted delivery of mesenchymal stem cells to infarct borders in Swine. Circulation 2003;108:2899–2904.[Abstract/Free Full Text]

39. Barbash IM, Chouraqui P, Baron J et al. Systemic delivery of bone marrow-derived mesenchymal stem cells to the infarcted myocardium: Feasibility, cell migration, and body distribution. Circulation 2003;108:863–868.[Abstract/Free Full Text]

40. Hofmann M, Wollert KC, Meyer GP et al. Monitoring of bone marrow cell homing into the infarcted human myocardium. Circulation 2005;111:2198–2202.[Abstract/Free Full Text]

41. Parish CR. Fluorescent dyes for lymphocyte migration and proliferation studies. Immunol Cell Biol 1999;77:499–508.[CrossRef][Medline]

42. Billinton N, Knight AW. Seeing the wood through the trees: A review of techniques for distinguishing green fluorescent protein from endogenous autofluorescence. Anal Biochem 2001;291:175–197.[CrossRef][Medline]

43. Kim S, Lim YT, Soltesz EG et al. Near-infrared fluorescent type II quantum dots for sentinel lymph node mapping. Nat Biotechnol 2004;22:93–97.[CrossRef][Medline]

44. Rieger S, Kulkarni RP, Darcy D et al. Quantum dots are powerful multipurpose vital labeling agents in zebrafish embryos. Dev Dyn 2005;234:670–681.[CrossRef][Medline]

45. Hanaki K, Momo A, Oku T et al. Semiconductor quantum dot/albumin complex is a long-life and highly photostable endosome marker. Biochem Biophys Res Commun 2003;302:496–501.[CrossRef][Medline]

46. Hsieh SC, Wang FF, Hung SC et al. The internalized CdSe/ZnS quantum dots impair the chondrogenesis of bone marrow mesenchymal stem cells. J Biomed Mater Res B Appl Biomater 2006;79:95–101.[Medline]