First published online in STEM CELLSExpress January 8, 2009.
Disclosure of potential conflicts of interest is found at the end of this article.
Author contributions: C.-C.W. and C.-H.C.: collection and assembly of data, data analysis and interpretation, manuscript writing; S.-M.H.: provision of study material, final approval of manuscript; W.-W.L., C.-H.H., and W.-Y.L.: collection and assembly of data, data analysis and interpretation; Y.C. and H.-W.S.: conception and design, manuscript writing, administration support, financial support, final approval of manuscript. C.-C.W. and C.-H.C. contributed equally to this article.
Cell transplantation via direct intramyocardial injection is a promising therapy for patients with myocardial infarction; however, retention of the transplanted cells at the injection sites remains a central issue following injection of dissociated cells. Using a thermoresponsive hydrogel system with a multiwell structure, we successfully developed an efficient technique to generate spherically symmetric bodies of mesenchymal stromal cells (MSCs) inherent with endogenous extracellular matrices (ECMs) for direct intramyocardial injection. After injection through a needle and upon transferring to another growth surface, the time required to attach, migrate, and proliferate was significantly shorter for the MSC bodies than the dissociated MSCs. Employing a syngeneic rat model with experimental myocardial infarction, an intramyocardial injection was conducted with a needle directly into the peri-infarct areas. There were four treatment groups (n = 10): sham, phosphate-buffered saline, dissociated MSCs, and MSC bodies. The results obtained in the echocardiography and catheterization measurements demonstrated that the MSC body group had a superior heart function to the dissociated MSC group. Histologically, it was found that MSC bodies could provide an adequate physical size to entrap into the interstices of muscular tissues and offer a favorable ECM environment to retain the transplanted cells intramuscularly. Additionally, transplantation of MSC bodies stimulated a significant increase in vascular density, thus improving the cardiac function. These results indicated that the spherically symmetric bodies of MSCs developed in the study may serve as a cell-delivery vehicle and improve the efficacy of therapeutic cell transplantation. STEM CELLS2009;27:724–732
Cell transplantation via direct intramyocardial injection is a promising therapy for patients with myocardial infarction [1, 2]. Prior to cell transplantation, a large scale of the desired cell types must be expanded in vitro on tissue culture polystyrene (TCPS) dishes. Upon confluence, detachment of the cultured cells from TCPS dishes often requires the use of a proteolytic enzyme. The use of enzyme commonly dissociates the cultured cells and disrupts their extracellular matrices (ECMs) and integrative adhesive agents .
Following injection of dissociated cells, retention of the transplanted cells at the injection sites remains a central issue [4, 5]. We speculated that dissociated cells injected into the heart may not have a sufficient physical size to entrap into the interstices of muscular tissues, thus causing the cell loss. It was reported that drug carriers with a diameter of 60-250 μm can be used for intramuscular injection . After injection, they are retained in the interstitial tissue acting as a sustained release depot [7, 8]. In this study, we hypothesized that the use of three-dimensional multicellular aggregates (cell bodies), with the preservation of endogenous ECMs, may significantly increase cell retention and is therefore beneficial for cell transplantation.
In our recent study, a cell culture system, employing a thermoresponsive methylcellulose (MC) hydrogel coated on TCPS dishes, was developed to cultivate human embryonic stem cell clumps for the formation of embryoid bodies . The cells within the embryoid bodies were shown to express molecular markers specific for representative cells from the three embryonic germ layers. However, the external morphology of thus obtained embryoid bodies was highly variable .
Bone marrow-derived mesenchymal stromal cells (MSCs) are thought to be multipotent cells  and have been applied in clinical trials, including the treatment of osteogenesis imperfecta [11, 12] and infarcted heart [13, 14]. Using a modified culture system, construction of spherically symmetric bodies of MSCs is reported in this study. Characteristics of MSC bodies, before and after injection through a needle, were examined in vitro. Additionally, direct intramyocardial injection of MSC bodies to induce angiogenesis and improve cardiac functions in a syngeneic rat model with infarcted myocardium was investigated. The dissociated MSCs obtained by traditional trypsinization were used as a control.
MATERIALS AND METHODS
Preparation of the Multiwell Cell Body Culture System
Aqueous MC solutions (12%, wt/vol) were prepared by dispensing the weighed MC powders (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) in heated water with the addition of phosphate-buffered saline (PBS, 5.0 g/l) at 50°C. The prepared MC solution was autoclaved and then kept in a refrigerator at 4°C for 24 hours. The obtained homogeneous MC solution was poured into a polystyrene tray (Nalge Nunc International, Rochester, NY, http://www.nalgenunc.com) and a 96-well amplification plate (Nalge Nunc International) was placed on top of it at 4°C (Fig. 1A). It was found that this specific aqueous MC underwent a sol-gel reversible transition upon heating or cooling at ∼32°C .
Subsequently, the tray was preincubated at 37°C for 2 hours and an opaque gel layer (3.3 ± 0.1 mm in thickness) with a multiwell structure (2.0 ± 0.1 mm in radius) was formed. After gelation, the 96-well amplification plate was removed and the obtained multiwell hydrogel system was used to cultivate cell bodies. The plain hydrogel system, without using the 96-well amplification plate to create the multiwell structure, was used as a control.
Cultivation of MSC Bodies
Bone marrow-derived MSCs were isolated from the femora and tibia of Lewis rats [16–18]. The isolated MSCs were spindle shaped and maintained mesenchymal differentiation potentials . To initiate cell differentiation, the DNA-demethylating agent 5-azacytidine (5-Aza, 10 μM; Sigma-Aldrich) was added on the 3rd day and incubated with MSCs for 24 hours . Subsequently, the induced MSCs were labeled for later identification by adding 100 μg/ml 5-bromo-2′-deoxyuridine (BrdU; Sigma-Aldrich) containing media to 70% confluent cultures for 24 hours . After reaching confluence, MSCs were dissociated from culture dishes with a 0.05% trypsin and then seeded in the prepared multiwell hydrogel system with a multichannel pipette at different cell densities (5.0 × 103, 1.0 × 104, 5.0 × 104, 1.0 × 105, or 2.0 × 105 cells per well) and cultured at 37°C for 24 hours to form cell bodies. It was demonstrated in our previous study that cells are not able to attach onto the surface of the MC hydrogel . Finally, the grown cell bodies were collected with a multichannel pipette and loaded in a syringe for cell transplantation (Fig. 1A).
Characterization of MSC Bodies
Photomicrographs of cell bodies grown in the multiwell hydrogel system were taken and their diameters were measured using Image-Pro Plus (version 4.5; Media Cybernetics, Bethesda, MD, http://www.mediacy.com) software (n = 10 batches). Examination of the morphology of cell bodies was performed with a scanning electron microscope (S-2300; Hitachi, Japan, http://www.hht-eu.com). The viability of cells in bodies was investigated according to a live/dead assay using calcein AM and ethidium homodimer (Invitrogen, Germany, http://www.invitrogen.com) . Additionally, cell bodies were trypsinized and subjected to trypan blue dye exclusion to determine total viable cells.
The cell morphology, endogenous ECMs, and integrative adhesive agents of MSC bodies, before and after injection through a needle, were examined. Briefly, MSC bodies (5 × 104 cells in total) were resuspended in 3 ml of culture medium, loaded in a syringe, injected through a 27-gauge needle, and subsequently seeded onto a 12-well culture plate (Costa Corning, Cambridge, MA, http://www.corning.com). Changes in morphology of MSC bodies on the plates with time were investigated and photographed. The dissociated MSCs (at the same cell density) were used as a control.
Paraformaldehyde-fixed MSC bodies were prepared for immunohistochemistry. The antibodies used were collagen type I (MP Biomedical, Solon, OH, http://www.mpbio.com), collagen type III (Chemicon, Temecula, CA, http://www.chemicon.com), fibronectin (Abcam, Cambridge, U.K., http://www.abcam.com), laminin (Chemicon), and E-cell adhesion molecule (E-CAM) (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com). Different Alexa Fluor secondary antibodies (Invitrogen) were used to obtain fluorescent colors. MSC bodies were costained to visualize F-actins and nuclei by phalloidin (Alexa Fluor 488 phalloidin) and propidium iodide (PI) (Sigma-Aldrich), respectively, and examined using an inverted confocal laser scanning microscope (CLSM) (TCS SL; Leica, Wetzlar, Germany, http://www.leica.com).
The investigation was conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication no. 85-23, revised in 1996) and approved by the Animal Care and Use Committee of Veterans General Hospital, Taichung, Taiwan. Acute myocardial infarction was created in male syngeneic Lewis rats weighing 300-350 g . After the permanent left coronary artery (LCA) ligation, color changes in the left ventricular (LV) muscle were notified in all rats . Thirty minutes after myocardial infarction, the rats were randomly divided into four treatment groups: sham (without the LCA ligation), PBS (300 μl), dissociated MSCs (5 × 105 cells) in PBS (300 μl), and MSC bodies (5 × 105 cells in total) in PBS (300 μl).
An intramyocardial injection of PBS, dissociated MSCs, or MSC bodies directly into the border zone of the infarct was performed with a 27-gauge needle [22, 23]. Animals were coded so that all measurements were made without knowledge of treatment groups. The study was continued until at least 10 rats survived for at least 3 months in each of the four coded groups.
LV Function Assessment by Echocardiography and Catheterization
Echocardiography was performed at 4, 8, and 12 weeks postoperatively for all studied groups. Rats were anesthetized with sodium pentobarbital (30 mg/kg), and isoflurane (2.0%) was used as a supplement to maintain mild anesthesia. Cardiac ultrasonography was carried out with a commercially available echocardiographic system (SONOS 5500; Philips Medical Systems, Bothell, WA, http://www.medical.philips.com) equipped with a 12-MHz broadband sector transducer. Dimension data were presented as the average of measurement of five consecutive beats. The fractional shortening of LV was calculated as follows :
where LVEDD and LVESD represent LV dimensions in end-diastole and end-systole, respectively.
Pressure measurements were performed at 12 weeks postoperatively . Rats were anesthetized with isoflurane (4.0%) and intubated for continuous ventilation with room air supplemented with oxygen and isoflurane (3.0%). The apex of LV was cannulated with a physiological pressure transducer (MLT844; Millar Instruments Inc., Houston, TX, http://www.millarinstruments.com). The pressure waveforms were recorded with a data-acquisition system (Powerlab ML870; AD Instruments, Milford, MA, http://www.adinstruments.com). The above-mentioned measurements were conducted by investigators blinded to the experimental conditions.
LV myocardium specimens were retrieved at day 1 (n = 5 for the dissociated MSC and MSC body groups only) or 12 weeks postoperatively (n = 10 for all studied groups). Specimens used for light microscopy were fixed in 10% phosphate-buffered formalin, embedded in paraffin, sectioned into a thickness of 5 μm, and then stained with Masson's trichrome. The stained sections were used to measure and calculate the thickness values of the peri-infarct and infarct areas in each studied group . The infarct size was expressed as the percentage of total LV circumference. Additional sections were stained for factor VIII with an immunohistological technique using the monoclonal anti-factor VIII antibody (DAKO, Glostrup, Denmark, http://www.dako.com) . The vascular density in the peri-infarcted area of all animals was quantified using the above-mentioned image analysis system.
For immunofluorescent staining, after rehydration and microwave antigen retrieval with 0.1 mol/l sodium citrate, sections were incubated at 4°C for 12 hours with the anti-BrdU antibody resuspended in the dilution buffer . The sections were then double-stained with antibodies against fibronectin, macrophage (CD68; AbD Serotec, Oxford, United Kingdom, http://www.ab-direct.com), connexin 43 (Cx43, Chemicon), α-sarcomeric actin (AbD Serotec), factor VIII, α-smooth muscle actin (α-SMA; DAKO), smooth muscle myosin heavy chain (SMMHC; Abcam), cleaved caspase-3 (Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com), α-actinin (Sigma-Aldrich), or the early marker of myocyte development Nkx2.5 (Santa Cruz Biotechnology Inc.). The stained sections were counterstained to visualize nuclei by Sytox blue (Invitrogen) or PI. The number of apoptotic cells (or macrophages infiltrated) per field, immunostained for cleaved caspase-3 (or CD68), was counted and expressed as a percentage of total cells.
Comparison between two groups was performed by the one-tailed Student's t test, and multiple group comparison was performed by one-way analysis of variance followed by Fisher's least significant difference using a statistical software (SPSS, Chicago, IL, http://www.spss.com). All data are presented as a mean value with its standard deviation indicated (mean ± SD). Differences were considered to be statistically significant when the p values were <.05.
Characteristics of MSC Bodies
The MSCs (84.5% ± 3.7% BrdU-labeled) seeded in the plain and multiwell hydrogel systems did not adhere onto substrates; instead, they aggregated and formed cell bodies with time. The morphology of MSC bodies formed in the plain hydrogel system was highly variable, whereas those generated in the multiwell hydrogel system were spherically symmetric (Fig. 1B). A single cell body was observed in each well in the multiwell hydrogel system, except for the case with a cell seeding density of 5 × 103 cells per well. The size of cell bodies grown in the multiwell hydrogel system increased significantly with increasing the cell seeding density (Table 1; supporting information Fig. 1). ECM molecules (collagen type I and type III), integrative adhesive agents (fibronectin and laminin), and intercellular junctions (E-CAM) were clearly identified (Fig. 2).
Table 1. Sizes of mesenchymal stromal cell bodies formed in the multiwell hydrogel system at different cell densities (n = 7 batches)
After injection through a 27-gauge needle (inside diameter: 400 μm), the MSC bodies formed at a cell seeding density of 1.0 × 104 cells per well (diameter: ∼195 μm) remained intact. In contrast, the bodies generated at a cell seeding density of 5.0 × 104 cells per well (diameter: ∼465 μm) or beyond often got stuck in the needle and were torn into pieces. Therefore, the cell bodies grown at a cell seeding density of 1.0 × 104 cells per well were chosen for further studies. Live/dead staining demonstrated that most of the cells in bodies were viable, based on the fluorescence images of 50 optical sections (Fig. 3). The total viable cells before and after injection (9,100 ± 85 and 8,900 ± 70 cells per body, respectively) were found to be comparable (p > .05; determined by trypan blue dye exclusion).
After injection, dissociated MSCs and MSC bodies were individually seeded onto 12-well plates. It took awhile for dissociated MSCs to settle down and spread out on the culture plate (Fig. 4A). Analyses of immunofluorescent images indicated that there was no fibronectin deposited on the plate surface initially. Six hours later, fibronectin was organized into short linear streaks and the cells started to attach to the plate surface (Fig. 4B).
In contrast, MSC bodies adhered to the culture plate shortly after seeding. Subsequently, the cells migrated out of bodies and attached and proliferated on the culture plate (Fig. 4A). A robust fibronectin meshwork inherent with the endogenous ECM was clearly observed in MSC bodies originally; this fibronectin meshwork started to attach to the plate surface within 1 hour and those cells which migrated out of bodies continuously produced fibronectin and deposited it onto the plate surface (Fig. 4B). The time required for cell confluence was shorter for the MSC body group (2-3 days) than the dissociated MSC group (4-5 days, p < .05).
The overall surgical mortality rate, defined as animal death within 24 hours after surgery, was 6.6% (four of 60 rats), and the late mortality rate (death between 24 hours and 12 weeks after surgery) was 8.9% (five of 56 rats) (PBS group, n = 3; dissociated MSC group, n = 1; MSC body group, n = 1).
LV Function Assessment by Echocardiography and Catheterization
The MSC body group showed a significantly greater LVFS when compared with the dissociated MSC group at 12 weeks postoperatively (Table 2; p < .05). The improvement in LV function for the group treated with MSC bodies was further indicated by a significant increase in LVESP and a decrease in LVEDP when compared with its counterpart treated with dissociated MSCs (Table 2; p < .05).
Table 2. Parameters of LV function and postmortem morphometry
Morphological and Histological Findings
A moderate degree in LV dilation and myocardial fibrosis was observed for the group treated with dissociated MSCs (supporting information Fig. 2). In contrast, the group treated with MSC bodies attenuated the enlargement of LV cavity and the development of myocardial fibrosis. The size of the infarct observed was significantly smaller in the MSC body group than in the dissociated MSC group, whereas its thickness values were significantly greater (Table 2; p < .05). The peri-infarct vascular density observed in the MSC body group (245 ± 19 vessels per mm2) was significantly greater than those seen in the PBS (95 ± 10 vessels per mm2; p < .05) and dissociated MSC (182 ± 11 vessels per mm2; p < .05) groups (Table 2).
At day 1 after intramyocardial injection, most of dissociated MSCs delivered to the heart through a needle were leaked back out of the injection site, but some were found in the myocardial interstices (Fig. 5A, 5B). In contrast, MSC bodies were able to entrap into the interstices of myocardial tissues and the transplanted cells were mostly localized at the site of injection (Fig. 5C). At 12 weeks postoperatively, there were a large number of BrdU-labeled cells adhered to fibronectin retained at the site of injection and there was little detectable cleaved caspase-3 in the MSC body group (<0.5%; Fig. 5D, 5E); however, only a few BrdU-labeled cells were identified in the dissociated MSC group. In the MSC body group, some neomicrovessel walls composed of BrdU-labeled endothelial cells (or smooth muscle cells [SMCs]) were recognized (Fig. 5F). Capillaries were identified as a single layer of factor VIII-positive cells with a flattened morphology, whereas arterioles were recognized as staining positive for α-SMA and as having a visible lumen.
A significant number of the BrdU-labeled cells were further stained positively for SMMHC (15.5% ± 2.6%; Fig. 5G) or α-SMA (13.3% ± 1.5%; Fig. 5H), indicating that a substantial portion of the implanted MSCs had been differentiated into SMCs or myofibroblasts. A few BrdU-labeled cells were stained positively for Nkx2.5 (Fig. 5I); however, no mature cardiomyocytes (α-actinin-positive cells) were identified. Also, little Cx43 was expressed at intercellular contact sites of the host/graft (Fig. 5J). However, it remained uncertain whether these characteristic features of transplanted cells were the result of transdifferentiation or fusion of cells. Quantification results demonstrated that the percentage of macrophages present at the site of intramuscular injection was 10.5% ± 2.8% at day 1 (Fig. 5K). At 12 weeks postoperatively, the number of macrophages decreased significantly (1.8% ± 0.6%; Fig. 5L).
Typical cell transplantation techniques involve the administration of dissociated cells by directly injecting them into the myocardium, in which the transplanted cells are not provided with a temporary matrix to which they can attach . In the study, we demonstrated that cell bodies can provide an adequate physical size to entrap into the myocardial interstices and offer a favorable ECM environment to retain the transplanted cells at the sites of injection.
It was shown in our previous study that the hydrated surface of the MC hydrogel is hydrophilic and neutrally charged . Such kind of culture surface can effectively inhibit the protein adsorption and the attachment of cells onto substrates . Previous work has shown that free-floating MSCs can form multicellular aggregates . Cell adhesion molecules such as integrins and cadherins have been implicated in participating in the process of formation of cell aggregates .
The cell bodies grown in the plain hydrogel system showed a variety of morphologies, as the free-floating MSCs adhered to each other in a random fashion in varying amounts. To overcome this problem, we seeded a fixed amount of cells in each well of the multiwell hydrogel system so that only the cells within each well could adhere to each other. This technique can produce spherically symmetric cell bodies with a relatively homogeneous size distribution in a short formation time (within 24 hours), factors that are crucial for a better control of cell delivery via intramuscular injection.
The MSC bodies grown at a cell seeding density of 1.0 × 104 cells per well had a radius of ∼100 μm, and most of the cells within bodies were viable as indicated by the live/dead staining assay. For the bodies generated at a cell seeding density of 5.0 × 104 cells per well or beyond (radius >200 μm), the cells deeply embedded inside the bodies were difficult to image by CLSM because of the penetration limit of the laser light (∼100 μm from the surface). Dense cellular structures develop hypoxia at distances beyond the diffusion capacity of oxygen (typically ∼200 μm in thickness). Beyond this thickness, the innermost cells are too far from the supply of oxygen and fresh growth medium to thrive . Therefore, it is likely that some cells in the interior of these cell bodies were hypoxic. It was reported that hypoxia might enhance the migratory capacity of MSCs [30, 31].
The obtained MSC bodies preserved the endogenous ECMs, which were constituted of proteins, such as collagen type I and type III, fibronectin, laminin, and E-CAM. After injection through a needle, we found that MSC bodies retained their activity upon transferring to another growth surface. Cell growth can be regulated by a number of ECM molecules including collagen and fibronectin . These matrix macromolecules are extremely useful for improving both cell adhesion and viability and controlling the host response that can then mediate cell attachment and spreading . In the study, we attempted to remove the cells from MSC bodies using a method descried in our previous study . However, we found that the structure of ECM bodies collapsed completely after removal of the cells from cell bodies. Therefore, we were not able to use ECM bodies as a control in the animal study.
At retrieval, only a few BrdU-labeled cells were found in the peri-infarcted area in the dissociated MSC group, whereas a large number of BrdU-labeled cells were identified in the MSC body group. This may be attributed to the fact that MSC bodies had a physical size larger than that of the dissociated MSCs and, therefore, had a better opportunity to entrap into the interstices of myocardial tissues. Once entrapped into the myocardium, the inherent ECM with MSC bodies could further provide a superior environment for the incorporation of transplanted cells to the host tissue.
It was reported that locally delivered MSCs were able to incorporate into newly formed vessels and displayed endothelial or SMC phenotype . Also, MSCs have been shown to express angiogenic growth factors in a paracrine fashion to stimulate neovascularization at the sites of the cell graft [20, 36]. These facts might explain why there was a significantly greater vascular density observed in the MSC body group than in the dissociated MSC group, consequently contributing toward an improved wall thickness and a reduction in the infarct size. Angiogenesis has been shown to contribute to the improvement of myocardial function by maintaining the viability of the grafted cells and residual cardiomyocytes [37, 38]. The results obtained in our echocardiography and catheterization measurements demonstrated that the MSC body group had a heart function superior to that of the dissociated MSC group.
Using a small animal model, a short-term proof-of-concept study showed the feasibility of our approach. However, a larger animal model in a long-term study would better simulate the conditions for patients with myocardial infarction. Additionally, further studies are required to determine the tumorigenicity of MSCs in cell bodies.
We successfully developed an efficient technique to generate spherically symmetric MSC bodies with the endogenous ECM for cellular cardiomyoplasty. The in vitro results demonstrated that their ability to attach to the underlying matrix and proceed to confluence was superior to that of dissociated MCSs. When injected into the peri-infarcted zone following experimentally induced myocardial infarction, there were significantly more cells retained in the MCS body group than in the dissociated MSC group, thus improving the cardiac function.
We thank I-Shou Chang for his help with statistical analyses. This work was supported by grants from the Veterans General Hospitals, the University System of Taiwan Joint Research Program (VGHUST96-P6-25), and the National Health Research Institute (NHRI-EX97-9518EI), Taiwan, Republic of China.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.