SEARCH

SEARCH BY CITATION

Keywords:

  • Mesenchymal stromal cells;
  • Mesenchymal stem cells;
  • Action potential;
  • Electrophysiology;
  • Ion channels;
  • Cell transdifferentiation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Although bone marrow-derived mesenchymal stromal cells (MSCs) may be beneficial in treating heart disease, their ability to transdifferentiate into functional cardiomyocytes remains unclear. Here, bone marrow-derived MSCs from adult female transgenic mice expressing green fluorescent protein (GFP) under the control of the cardiac-specific α-myosin heavy chain promoter were cocultured with male rat embryonic cardiomyocytes (rCMs) for 5–15 days. After 5 days in coculture, 6.3% of MSCs became GFP+ and stained positively for the sarcomeric proteins troponin I and α-actinin. The mRNA expression for selected cardiac-specific genes (atrial natriuretic factor, Nkx2.5, and α-cardiac actin) in MSCs peaked after 5 days in coculture and declined thereafter. Despite clear evidence for the expression of cardiac genes, GFP+ MSCs did not generate action potentials or display ionic currents typical of cardiomyocytes, suggesting retention of a stromal cell phenotype. Detailed immunophenotyping of GFP+ MSCs demonstrated expression of all antigens used to characterize MSCs, as well as the acquisition of additional markers of cardiomyocytes with the phenotype CD45-CD34+-CD73+-CD105+-CD90+-CD44+-SDF1+-CD134L+-collagen type IV+-vimentin+-troponin T+-troponin I+-α-actinin+-connexin 43+. Although cell fusion between rCMs and MSCs was detectable, the very low frequency (0.7%) could not account for the phenotype of the GFP+ MSCs. In conclusion, we have identified an MSC population displaying plasticity toward the cardiomyocyte lineage while retaining mesenchymal stromal cell properties, including a nonexcitable electrophysiological phenotype. The demonstration of an MSC population coexpressing cardiac and stromal cell markers may explain conflicting results in the literature and indicates the need to better understand the effects of MSCs on myocardial injury.

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


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Author contributions: R.A.R. and H.J.: conception and design, provision of study materials, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; R.A.R. and H.J. contributed equally to this work. X.W., S.H., and J.N.T.: conception and design, provision of study materials, collection and assembly of data, data analysis and interpretation, final approval of manuscript; N.G.: conception and design, final approval of manuscript; S.C.J.K.: collection and assembly of data, data analysis and interpretation, final approval of manuscript; T.G.P.: conception and design, financial support, provision of study materials, data analysis and interpretation, final approval of manuscript; P.H.B. and A.K.: conception and design, financial support, provision of study materials, data analysis and interpretation, manuscript writing, final approval of manuscript.

As a result of the limited ability of the injured heart to regenerate cardiomyocytes, cell-based therapies have been advocated as a possible treatment for heart failure. Although several types of progenitor cells may be appropriate for cell-based tissue repair in patients with heart failure [1, [2]3], multipotent mesenchymal stromal cells (MSCs) from the bone marrow are particularly attractive because they can easily be extracted and expanded in culture [4, [5], [6], [7]8]. Several reports claim that bone marrow-derived MSCs adopt a cardiomyocyte phenotype both in vitro and when injected into healthy or infarcted hearts [9, [10], [11], [12], [13], [14], [15], [16]17]. These studies were based mainly on the expression of cardiac markers such as natriuretic peptides and sarcomeric proteins, although one study [11] concluded that MSCs produce sarcomeres and generate action potentials after chronic treatment with the DNA-demethylating agent 5-azacytidine. Other studies, however, concluded that MSCs do not transdifferentiate into functional cardiomyocytes, suggesting that any beneficial effects of MSCs might occur through indirect mechanisms such as angiogenesis or the release of paracrine factors [18, [19], [20], [21], [22], [23]24].

The goal of this study was to directly test whether adult bone marrow-derived MSCs can transdifferentiate into functional cardiomyocytes using a novel in vitro system. MSCs were isolated from transgenic mice expressing the green florescent protein (GFP) under the control of the cardiac-specific α-myosin heavy chain promoter and cocultured with rat embryonic cardiomyocytes (rCMs). With this system, mouse MSCs that acquire a cardiomyocyte phenotype are detectable by the expression of GFP (GFP+). Our results demonstrate that although several cardiac markers were detected in GFP+ MSCs, these cells did not display the typical electrical properties of true cardiomyocytes. Moreover, the GFP+ MSCs continued to express all the markers used to characterize murine MSCs, indicating plasticity toward the cardiomyocyte lineage rather than true cardiac transdifferentiation under these specific in vitro conditions. Some of these data have been presented in abstract form [25].

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

B-a-Fvb Transgenic Mice

A transgenic vector consisting of a GFP coding sequence flanked by the full-length mouse α-myosin heavy chain (α-MHC) promoter was kindly provided by Dr. Takayuki Morisaki (National Cardiovascular Center, Osaka, Japan). A 5.5-kilobase promoter sequence was excised by BamHI and ScalI and subcloned into pBluescriptIIKS (Stratagene, La Jolla, CA, http://www.stratagene.com) to generate pKS-MHC. A PGFP-N1 plasmid (Clontech, Palo Alto, CA, http://www.clontech.com) was used to excise a 0.76-kilobase GFP full coding sequence using ScalI and NotI. The NotI restriction enzyme site was blunted and linked with HindIII. This fragment was then inserted into ScalI- and HindIII-digested pKS-MHC, resulting in a newly constructed plasmid was that was designated pNC26-α-MHC-GFP. The transgene sequence was obtained by using NotI to remove the pBluescript vector sequence from the pNC26-α-MHC-GFP construct. This purified fragment was microinjected into the fertilized oocytes of Fvb mice according to standard protocols. Founder animals were produced and mated with wild-type Fvb mice, and transgenic offspring were identified by polymerase chain reaction (PCR) genotyping of unique transgene sequences from tail genomic DNA. These transgenic mice were designated B-a-Fvb mice. All experiments were performed with the transgenic B-a-Fvb mice and wild-type littermates as controls.

Bone Marrow-Derived Mesenchymal Stromal Cells

Bone marrow MSCs were collected from the femur and tibia of individual adult female B-a-Fvb mice. These MSCs were cultured in long-term marrow culture medium according to our previously described method [26]. Briefly, 4–5 × 107 MSCs per mouse were collected and cultured using a long-term marrow culture medium consisting of 12.5% fetal bovine serum, 12.5% horse serum, 1% antibiotic-antimycotic solution, 1% glutamine, 1% vitamin solution, 0.8% essential amino acid, 0.4% nonessential amino acid, 1% sodium pyruvate, 1% sodium bicarbonate solution, 0.036% hydrocortisone (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 0.5 ng/ml interleukin-1α (R&D Systems Inc., Minneapolis, http://www.rndsystems.com), and 0.1 ng (25 units)/ml fibroblast growth factor. Unless otherwise specified, all reagents were from Gibco (Grand Island, NY, http://www.invitrogen.com). After 5–7 days in culture, half of the supernatant was replaced with fresh medium. At 2 weeks, when the adherent layer became confluent, the cells were detached by treating them with 0.25% trypsin and 1 mM EDTA-4Na, and after inactivation with 20% serum and washing with phosphate-buffered saline, the cells were resuspended and serially passaged when they reached near confluence (every 5–7 days) until fourth passage (T4) cells were obtained. The cells were incubated at 37°C in a humidified incubator containing 5% CO2. On the basis of detailed flow cytometry, immunophenotyping, and lineage differentiation experiments, the adherent B-a-Fvb MSC population displayed the properties consistent with the requirements for the definition of MSCs [6, 27] as described in detail in Results.

Coculture of B-a-Fvb Bone Marrow MSCs with Rat Embryonic Cardiomyocytes

Pregnant rats (day 15) were sacrificed by CO2 inhalation, the embryos were quickly excised, and the hearts were dissected out. Cardiomyocytes were isolated and cultured on the basis of the procedures previously described for rat neonatal cardiomyocytes [28, 29]. Only male rat embryonic cardiomyocytes were used in the coculture experiments. After 24 hours in culture, fourth passage bone marrow-derived MSCs from female B-a-Fvb mice were plated onto the primary rat embryonic cardiomyocyte cultures and cocultured for 5–15 days (supplemental online Fig. I). This time frame for coculture was chosen because it takes approximately 7–10 days for spontaneously active cardiomyocytes to become detectable in developing embryos or embryoid bodies [30]. Furthermore, this time course is comparable to that of other studies that have reported transdifferentiation of mouse MSCs toward the cardiomyocyte lineage [31]. MSCs were also cocultured with isolated embryonic rat lung, kidney, and liver cells.

mRNA Measurements

Northern blotting was used to detect GFP by extracting tissue RNA using a Trizol reagent (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) according to the instructions of the manufacturer. RNA (15 μg) was electrophoresed and transferred to a Hybond N+ membrane (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com). Digoxigenin-labeled GFP probes were hybridized at 50°C, and a DIG luminescent detection kit (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com) was used for final detection of RNA.

RNA was also analyzed using ThermoScript (Invitrogen) one-step quantitative reverse transcription (RT)-PCR as described previously [32]. Specifically, a platinum Taq kit (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) was used to synthesize the cDNA. Gene-specific sequences of mouse-specific oligonucleotide primers for Nkx2 transcription factor-related locus 5 (Nkx2.5), atrial natriuretic factor (ANF), and α-cardiac actin were based on DNA sequences in the National Center for Biotechnology Information. The internal standard primer was glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Mouse-specific primers were as follows: Nkx2.5 sense, 5′-ATGCCTATGGCTACAACGC-3′; Nkx2.5 antisense, 5′-ACTCACTTTAATGGGAAGAGGG-3′; ANF sense, 5′-TTGGCTTCCAGGCCATAATTG-3′; ANF antisense, 5′-AAGAGGGCAGATCTATCGGA-3′; α-cardiac actin sense, 5′-TGTTACGTCGCCTTGGATTTTGAG-3′; α-cardiac actin antisense, 5′-AAGAGAGAGACATATCAGAAGC)-3′; GAPDH sense, 5′-TTCTTGTGCAGTGCCAGCCTCGTC-3′; and GAPDH antisense, 5′-TAGGAACAGGGAAGGCCATGCCAG-3′.

The gene-specific oligonucleotide primers were added to a mixture containing a 1× concentration of SYBR Green PCR master mix, 0.25 U/ml multiscribe reverse transcriptase, 0.4 U/ml RNase inhibitor, and 10 ng of total tissue/cellular RNA in a 50-μl mixture according to the manufacturer's protocol (Applied Biosystems). The PCR protocol included an initial 30 minutes at 48°C (for cDNA synthesis) and then denaturation at 95°C for 10 minutes, followed by 40 cycles of denaturation at 95°C for 15 and 60 seconds at 60°C for annealing and elongation. Optics were on for fluorescence monitoring. The specificity and purity of the amplification reaction was determined by performing a melting curve analysis. The relative quantification of gene expression by real-time RT-PCR in a sample was determined by comparing the target-amplified product against GAPDH (internal standard) within the same sample. GAPDH mRNA expression was not significantly different among the various groups. Data were normalized to GAPDH expression and are shown as fold increases relative to gene expression in adult mouse heart [32].

Immunohistochemistry

Immunohistochemical techniques were used to detect GFP, cardiac troponin-I (cTnI), α-sarcomeric actinin, and collagen IV in B-a-Fvb bone marrow MSCs cocultured with rat embryonic cardiomyocytes. The following antibodies were used for immunohistochemical studies. For GFP, a rabbit anti-GFP polyclonal antibody was used, whereas the secondary antibody was an anti-rabbit IgG-linked fluorescein isothiocyanate (Molecular Probes, Eugene, OR, http://probes.invitrogen.com). For cTnI, a goat anti-mouse cTnI polyclonal antibody was used, along with an anti-goat IgG-linked Alexa 555 secondary antibody. For α-sarcomeric actinin, a monoclonal anti-sarcomeric actinin antibody (Sigma-Aldrich) was used with an Alexa Fluor 555 anti-mouse IgG (H+L) secondary antibody. For collagen type IV, a rabbit anti-mouse collagen type IV polyclonal antibody (BioDesign, Saco, ME, http://meridianlifescience.com) was used in conjunction with an anti-rabbit IgG-linked Alexa 555 secondary antibody (Molecular Probes). Nuclei were stained using 4,6-diamidino-2-phenylindole (DAPI), which was used to quantify cell numbers. Detection of fluorescent signals for each of these markers was performed using confocal laser scanning microscopy (Zeiss Axioplan 2 upright microscope with LSM 510 laser scanning module and Axiovision software; Carl Zeiss, Jena, Germany, http://www.zeiss.com).

Enumeration of Cell Fusion

For cell fusion studies, rCMs from male rats were cultured, as described above, and plated in eight-well culture dishes for 24 hours. Following this, fourth passage B-a-Fvb MSCs from female mice were cocultured with the rCMs for 5 days. On day 5 of coculture, cells were fixed, and fluorescence in situ hybridization was used to detect X and Y chromosomes using custom-made probes (rat Y chromosome-Cy 5-streptavidin biotin conjugate and mouse X chromosome-Cy3; Cambio, Dry Drayton, U.K., http://www.cambio.co.uk). Because the Y chromosome probe was rat-specific and the X chromosome probe was mouse-specific, visualization of cells with two X chromosomes (from female mice) and one Y chromosome (from male rats) were taken to have undergone a fusion event (XXY chromosomes).

Cell Sorting and Immunophenotyping

Bone marrow-derived MSCs from B-a-Fvb mice were cocultured with rCMs for 5, 7, and 10 days as described above (also described in supplemental online Fig. I), and stained with antibody-linked PE against a panel of cell surface antigens and proteins characteristically found in either bone marrow MSCs or cardiomyocytes (Table 1; supplemental online Table 1). Cell sorting was performed using BD FACSAria. GFP detection was with a 530/30 nM band-pass filter and a 505 nM long-pass filter. Sorting was done at a pressure of 20 psi with a 100-μm nozzle. The sorted GFP+ cells were centrifuged (1,000 rpm, 10 minutes) and spun down on gelatin-coated slides using Thermo Shandon (Cytospin 4) machinery (Thermo Shandon Inc., Pittsburgh, http://www.thermo.com). The cells were fixed with 4% paraformaldehyde and stained with the antibodies described in supplemental online Table 1. Samples were observed by confocal laser scanning microscopy (Zeiss Axioplan 2 upright microscope with LSM 510 laser scanning module and Axiovision software).

Table Table 1.. Immunophenotyping of B-a-Fvb bone marrow MSCs before and after coculturing with rCMs
  1. aB-a-Fvb bone marrow MSCs cultured alone an analyzed after fourth passage.

  2. bB-a-Fvb bone marrow MSCs were cocultured with rCMs and the GFP+ MSCs were sorted as described in Materials and Methods.

  3. c+, 5%–9% expression; ++, 10%–69% expression; +++, more than 70% expression; ++++, more than 95% expression.

  4. Abbreviations: GFP, green fluorescent protein; MSC, mesenchymal stromal cell; rCM, rat embryonic cardiomyocyte.

Thumbnail image of

Electrophysiology

Electrophysiological measurements were performed on GFP rCMs in coculture and cultured alone, GFP+ MSCs, and bone marrow-derived MSCs cultured alone. Recordings were performed at culture/coculture time points of 5, 7, 10, and 15 days. GFP fluorescence was observed following excitation with a mercury lamp. Current-clamp recordings were performed using the perforated patch clamp technique [33], whereas voltage-clamp measurements were made using the whole-cell configuration of the patch clamp technique [34].

Some rCMs exhibited spontaneous, rhythmic electrical activity that was recorded in current-clamp mode. In cells that did not display spontaneous activity, action potentials were evoked by applying 300–600 pA depolarizing current pulses for 4 milliseconds. To record sodium current (INa) and calcium current (ICa), cells were held at −100 mV and then given 250-millisecond voltage clamp steps to −50 and 0 mV, to activate INa and ICa, respectively. Cells were also studied using voltage ramps from −100 to +100 mV over 1 second.

For recording action potentials, the recording chamber was superfused with a normal Tyrode's solution (22°C–23°C) containing 140 mM NaCl, 5.4 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM HEPES, and 5 mM glucose, with pH adjusted to 7.4 with NaOH. To record INa and ICa, the following external solution was used: 140 mM NaCl, 5.4 mM TEA-Cl, 1 mM MgCl2, 2 mM CaCl2, 10 mM HEPES, and 5 mM glucose, with pH adjusted to 7.4 with CsOH. The pipette solution for recording action potentials consisted of 135 mM KCl, 5 mM NaCl, 0.1 mM CaCl2, 10 mM EGTA, 4 mM magnesium-ATP, 1 mM MgCl2, 6.6 mM sodium-phosphocreatine, 0.3 mM sodium-GTP, and 10 mM HEPES, with pH adjusted to 7.2 with KOH. Amphotericin B (Sigma-Aldrich) was included in the patch pipette at a concentration of 200 μg/ml to establish the perforated patch configuration. For INa and ICa measurements, the pipette solution contained 135 mM CsCl, 5 mM NaCl, 0.1 mM CaCl2, 10 mM EGTA, 4 mM magnesium-ATP, 1 mM MgCl2, 6.6 mM sodium-phosphocreatine, 0.3 mM sodium-GTP, and 10 mM HEPES, with pH adjusted to 7.2 with CsOH.

Micropipettes were pulled from borosilicate glass (with filament, 1.5 mm OD, 0.75 mm ID; Sutter Instrument, Novato, CA, http://www.sutter.com) using a Flaming/Brown pipette puller (model p-87; Sutter Instrument). The resistance of these pipettes was 5–8 MΩ when filled with recording solution and fire-polished. Microelectrodes were positioned with a micromanipulator (Burleigh PCS-5000 system, Mississauga, Ontario, http://www.exfo-burleigh.com) mounted on the stage of an inverted microscope (Olympus IX51; Olympus, Tokyo, http://www.olympus-global.com). Seal resistances were 2–20 GΩ. For action potential recording with the perforated patch clamp technique, access resistance was monitored using an Axopatch 200B amplifier (Axon Instruments, Union City, CA, http://www.moleculardevices.com). An access resistance of less than 40 MΩ, which generally developed within 5 minutes, was sufficient to record membrane potential. For voltage clamp experiments in the whole-cell configuration, the sarcolemma in the patch was ruptured, which resulted in access resistances of 5–15 MΩ. Data were digitized using a Digidata 1322A interfaced with pCLAMP 9 software (Axon Instruments). Data were stored on the computer for analysis offline.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Positive GFP fluorescence was observed in the hearts of B-a-Fvb transgenic mice expressing GFP under the control of the cardiac-specific α-myosin heavy chain promoter but not in wild-type mice (supplemental online Fig. II). Northern blot analysis established that GFP was highly expressed in the hearts of B-a-Fvb mice but not in any other tissues, including spleen, liver, lung, skeletal muscle, kidneys, or bone marrow (supplemental online Fig. II).

We next confirmed that MSCs isolated and cultured from these B-a-Fvb mice displayed the necessary properties to be classified as true MSCs [6, 27]. This was done using both immunostaining (Table 1) and flow cytometry analysis (supplemental online Fig. III). As expected, the adherent B-a-Fvb MSC population expressed the immunophenotype consistent with the accepted definition of MSCs: namely CD73+, CD90+, CD105+, CD45 CD11b, CD19, and CD79. Importantly, the absence of CD45-expressing cells by flow cytometry analysis (supplemental online Fig. III) confirms that the MSC cultures did not contain cells from the hematopoietic lineage, suggesting that our MSCs were not contaminated by hematopoietic stem/progenitor cells. In addition, these B-a-Fvb MSCs were observed to differentiate into osteoblasts and adipocytes in vitro when appropriate conditioned medium for each lineage was used [35, 36], as expected for MSCs (data not shown).

To examine whether MSCs could transdifferentiate into cardiomyocytes, bone marrow-derived MSCs from B-a-Fvb mice were cocultured with rCMs. After 5 days in coculture, 6.3% ± 2.1% of the MSCs showed positive GFP fluorescence (GFP+; supplemental online Fig. IV; representative images of GFP+ MSCs are given in Figs. 2, 6). Cell number was quantified using DAPI nuclear staining. In contrast, no GFP+ cells were observed when B-a-Fvb MSCs were cultured alone for up to 20 passages (not shown) or when wild-type Fvb MSCs were cocultured with rCMs (supplemental online Fig. IV). In addition, B-a-Fvb MSCs cocultured with rat embryonic kidney, liver, or lung did not yield any GFP+ cells (supplemental online Fig. V).

To assess whether other cardiac-specific promoters were activated in the cocultured MSCs, we performed quantitative PCR using mouse-specific primers to avoid detection from rCMs. After 5, 7, and 10 days in coculture, MSCs expressed ANF, Nkx2.5, and α-cardiac actin, whereas the expression of these genes was nominally zero in MSCs cultured alone. Surprisingly, the expression levels were highest at day 5 and decreased steadily thereafter (Fig. 1). The expression of these genes was undetectable in adult rat heart or in rCMs cultured alone, confirming the specificity of our mouse primers (data not shown). Consistent with activation of the cardiomyocyte genetic program, immunohistochemical studies revealed that GFP+ MSCs in coculture for 5 days also stained positively for the mouse cardiac-specific sarcomeric proteins cTnI and α-actinin (Fig. 2A, 2B).

thumbnail image

Figure Figure 1.. mRNA expression in bone marrow-derived MSCs cocultured with rat embryonic cardiomyocytes for 5, 7, and 10 days. Mouse-specific polymerase chain reaction primers were used to detect the expression of ANF (A), Nkx2.5 (B), and α-cardiac actin (C), which were normalized to glyceraldehyde-3-phosphate dehydrogenase expression. Data are presented as fold increases relative to the expression of these genes in adult mouse heart, which was normalized to 1. Data are mean ± SE based on 10 separate experiments for each condition. *, p < .05 compared with adult mouse heart; +, p < .001 compared with expression in B-a-Fvb MSCs; #, p < .05 compared with respective gene expression at 10 days of coculture. x-Axis and coculture time point labels apply to all panels. Abbreviations: ANF, atrial natriuretic factor; MSC, mesenchymal stromal cell; rCM, rat embryonic cardiomyocyte.

Download figure to PowerPoint

thumbnail image

Figure Figure 2.. Immunohistochemical analysis of green fluorescent protein (GFP)+ cells in cocultures containing rat embryonic cardiomyocytes and B-a-Fvb mesenchymal stromal cells. (A): Detection of mouse cardiac troponin-I (cTnI). Representative images are presented for mouse GFP (panel 1), cTnI (panel 2), the overlay of these two images (panel 3), and the corresponding bright-field image (panel 4). (B): Detection of mouse α-sarcomeric actinin. Representative images are presented for GFP (panel 1), mouse α-sarcomeric actinin (panel 2), the overlay of these two images (panel 3), and the corresponding bright-field image (panel 4). Note the coexpression of cTnI and α-sarcomeric actinin in the GFP+ cells in these cocultures.

Download figure to PowerPoint

We next investigated the possibility that cell fusion between rCMs (from male rats) and MSCs (from female mice) could be responsible for expression of GFP by using fluorescence in situ hybridization to detect X and Y chromosomes in the cells. Fusion events (as evidenced by the detection of cells with XXY chromosomes; described in Materials and Methods) did occur, but at the very low frequency of 0.68% (8 of 1,177 cells), suggesting that this is not a major factor in our studies.

To more definitively assess the ability of MSCs to develop a cardiomyocyte phenotype, electrophysiological recordings were performed on GFP+ MSCs after 5–15 days in coculture. Figure 3 shows that spontaneous and evoked action potentials (APs) were routinely observed when current-clamp recordings were made in cocultured rCMs (GFP) or in rCMs cultured alone. When rCMs displayed spontaneous APs, distinct diastolic depolarization phases were observed, with the maximum diastolic potential typically reaching approximately −65 mV, similar to myocytes from the cardiac conduction system [37]. Those rCMs not showing spontaneous APs had stable resting membrane potentials of ∼−75 mV, typical of atrial or ventricular myocytes, and in these myocytes APs could be evoked with small current injections. Consistent with the presence of APs in rCMs, Figure 4A shows that voltage-gated INa and ICa currents could readily be recorded using standard voltage-clamp protocols after 5 days in coculture. Similar findings were observed at the other coculture time points.

thumbnail image

Figure Figure 3.. Action potential recordings from rat embryonic cardiomyocytes (rCMs) cocultured with B-a-Fvb mesenchymal stromal cells. (A): Representative recording showing spontaneous action potentials observed in a green fluorescent protein (GFP) myocyte. Maximum diastolic potential was approximately −65 mV. (B): Representative recording of an action potential elicited with a 300-pA depolarizing pulse in a quiescent GFP myocyte. Resting membrane potential was approximately −75 mV. Dashed line indicates the 0-mV level. Data are representative of measurements made on nine GFP rCMs. Action potentials could not be elicited in any GFP+ MSCs in coculture with rCMs (n = 10) or in bone marrow MSCs cultured alone. Measurements were performed after 5 days of coculture. Abbreviations: ms, millisecond; s, second.

Download figure to PowerPoint

thumbnail image

Figure Figure 4.. Measurement of voltage-gated INa and ICa. (A): Representative recordings during voltage clamp steps to −50 and 0 mV are presented for a GFP rCM in coculture, a GFP+ MSC in coculture, and an MSC cultured alone. Voltage clamp protocol is displayed at bottom of (A). INa and ICa are expressed and were recorded in all myocytes at −50 and 0 mV, respectively. In contrast, these ionic currents were not detected in any GFP+ MSCs in coculture with rCMs or in bone marrow stromal cells cultured alone. Summary data are presented in (B) (mean ± SE; n = 14 GFP rCMs; n = 24 GFP+ MSCs; n = 10 GFP MSCs). Data shown are from day 5 cocultures. Abbreviations: GFP, green fluorescent protein; ICa, calcium current; INa, sodium current; ms, millisecond; MSC, mesenchymal stromal cell; ND, not detectable; rCM, rat embryonic cardiomyocyte.

Download figure to PowerPoint

In contrast to the findings in rCMs, cocultured GFP+ MSCs did not display spontaneous rhythmic activity, nor could APs be evoked. Consistent with this lack of excitability, the voltage-clamp protocols used to identify INa or ICa in rCMs did not generate these currents in either the GFP+ MSCs in coculture or the MSCs cultured alone (Fig. 4). The data shown in Figure 4 are from day 5 cocultures; identical responses were measured in day 7, 10, and 15 cocultures (supplemental online Fig. VI).

To further characterize the electrical behavior of our cultured cells, the responses of rCMs, as well as GFP+ MSCs in coculture and MSCs cultured alone, were recorded in response to slow (1 second) voltage ramps from −100 to +100 mV (Fig. 5). Although voltage-gated inward currents were clearly activated in rCMs using this protocol, GFP+ MSCs displayed only small, outwardly rectifying currents, typical of nonexcitable cells [38]. Similar results were obtained in GFP+ MSCs at 5, 7, 10, and 15 days of coculture (supplemental online Fig. VI). The currents in cocultured GFP+ MSCs were not different (p = .29) from those of MSCs cultured alone at any of the time points investigated (Fig. 5C).

thumbnail image

Figure Figure 5.. Current responses during voltage ramps in rCMs cocultured with bone marrow-derived MSCs for 5–15 days. (A): Representative recordings are illustrated for a rCM in coculture, a GFP+ MSC in coculture, and an MSC cultured alone during a voltage ramp from −100 to + 100 mV ([A], bottom). (B): The current responses from (A) are shown together as a function of voltage. Note the presence of the rapidly activating and inactivating Na+ current in the rCM. The GFP+ MSC and the normal MSC both show only a weakly outwardly rectifying response, with no voltage-dependent inward currents. Representative traces in (A) and (B) are from day 5 cocultures. (C): Summary data (mean ± SE) showing maximum inward and outward current densities, measured at +100 and −100 mV, for GFP+ MSCs in coculture and normal MSCs cultured alone for 5, 7, 10, and 15 days. Note that the response in GFP+ MSCs is very similar to that of normal MSCs at all time points. For GFP+ MSCs, n = 18 at day 5, 9 at day 7, 9 at day 10, and 8 at day 15. For MSCs, n = 10 at day 5, 7 at day 7, 7 at day 10, and 5 at day 15. Abbreviations: GFP, green fluorescent protein; ms, millisecond; MSC, mesenchymal stromal cell; rCM, rat embryonic cardiomyocyte.

Download figure to PowerPoint

The results above establish that the electrophysiological characteristics of GPF+ MSCs are distinct from those of cardiomyocytes and identical to those of nonexcitable MSCs, suggesting that GPF+ MSCs retained a stromal phenotype under these in vitro conditions. Indeed, immunohistochemical analysis established that GFP+ MSCs express collagen type IV (Fig. 6), a marker for stromal cells that is absent on embryonic cardiomyocytes [39]. To further investigate whether GFP+ MSCs retain a stromal phenotype despite clear evidence for the expression of several cardiac genes, we performed additional immunophenotyping on sorted GFP+ MSCs from B-a-Fvb MSC/rCM cocultures (described in Materials and Methods). As shown in Table 1 (and supplemental online Fig. III), B-a-Fvb MSCs cultured alone expressed CD73, CD90, CD105, collagen type IV, and vimentin, but not CD45, CD11b, CD19, CD34, CD79, troponin T, troponin I, α-actinin, or OX40L (CD134L), as expected for MSCs [6, 27]. After 5, 7, and 10 days of coculture, the GFP+ MSCs continued to express these MSC markers at all time points, but in addition, cardiac troponin T, troponin I, α-actinin, connexin 43, OX40L, and CD34 were expressed (Table 1). Together, these data establish that under these experimental conditions the GFP+ cells, although they express some cardiac proteins, continue to express the full complement of surface antigens used to define MSCs.

thumbnail image

Figure Figure 6.. Green fluorescent protein (GFP)+ mesenchymal stromal cells express collagen type IV. Representative images are presented for GFP (panel 1), mouse collagen IV (panel 2), and the overlay of these two images (panel 3), which illustrates that the GFP+ MSCs express mouse collagen type IV, a stromal cell marker not found in cardiomyocytes. The corresponding bright-field image is shown in panel 4.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

We have developed an in vitro model that assesses the ability of adult bone marrow-derived MSCs to transdifferentiate and acquire a cardiomyocyte-like phenotype under conditions of coculture with rCMs. Specifically, MSCs were obtained from transgenic mice (B-a-Fvb), which express GFP under the control of the α-myosin heavy chain promoter. Therefore, cells undergoing transdifferentiation toward the cardiomyocyte lineage are expected to express GFP. We found that coculturing of B-a-Fvb MSCs with rCMs resulted in a significant frequency of GFP+ cells (6.3%), whereas no GFP+ cells were detected in B-a-Fvb MSCs when cultured alone or when cocultured with noncardiac embryonic tissue. In further assessing the transdifferentiation properties of cocultured GFP+ MSCs using quantitative PCR, we found that cocultured GFP+ MSCs expressed several cardiac markers, such as ANF, Nkx2.5, and α-cardiac actin. In contrast, these cardiac genes were undetectable in MSCs cultured alone. By validating that the primers used in our quantitative PCR measurements were mouse-specific, we conclude that the cardiac genetic program is activated in MSCs placed in a milieu generated by the embryonic cardiomyocytes. Interestingly, cardiac gene expression was greatest at 5 days of coculture and decreased steadily thereafter. Although the basis for these time-dependent reductions in cardiac gene expression is not clear, this reduction suggests that the MSC phenotype is highly dynamic, which could limit plasticity of MSCs as discussed below.

To further assess whether GFP+ MSCs transdifferentiated toward a cardiac phenotype, we performed patch-clamp electrophysiological recordings to explore whether GFP+ MSCs had electrical properties typically seen in excitable cells like cardiomyocytes [37, 40]. Unlike rCMs, which generated spontaneous or evoked APs and produced INa and ICa typical of cardiomyocytes; APs, INa, and ICa were not observed in GFP+ MSCs when cocultured with rCMs or when cultured alone. These results are consistent with previous studies showing that INa and ICa are either absent or very small in MSCs [41, 42]. On the other hand, cocultured GFP+ MSCs and MSCs cultured alone generate outwardly rectifying currents, which are characteristic of nonexcitable cells such as cardiac fibroblasts [38]. In this regard, it is interesting that previous studies have suggested that MSCs acquire a fibroblast-like appearance when injected into rats with myocardial infarction [15]. Thus, our electrical studies establish that under these specific in vitro conditions, GFP+ MSCs do not undergo electrical transdifferentiation toward a cardiomyocyte phenotype even when cocultured for up to 15 days with rCMs, despite evidence for the expression of some cardiac-specific genes. Rather, the GFP+ MSCs were more akin to the nonexcitable MSCs from which they are derived. This is in contrast to studies on other types of progenitor cells, such as embryonic stem cells [30] and resident cardiac stem cells [43, 44], which may produce cells with electrical properties resembling cardiomyocytes under in vitro culture conditions.

Since cocultured GFP+ MSCs showed electrical properties more closely resembling those of unexcitable mesenchymal stromal cells, we undertook further extensive immunophenotypic analysis of cocultured GFP+ MSCs. Immunostaining of the GFP+ MSCs revealed the retention of the standard MSC markers (CD73, CD90, CD105, SDF1, CD44, CD34, collagen type IV, vimentin) and a lack of CD45 [6]. In addition to CD34, GFP+ MSCs acquired determinants associated with cardiomyocytes, such as troponin T, troponin I, α-actinin, and connexin 43, as well as CD134L, which is the ligand for Ox40, a member of the tumor necrosis factor superfamily associated with immune activation also expressed on cardiomyocytes under some circumstances [45]. Thus, the GFP+ MSCs (the “putative cardiomyocytes”) are in fact mesenchymal stromal cells with the immunophenotype CD45, CD73+, CD90+, CD105+, SDF1+, CD44+, CD34+, CD134L+, collagen type IV+, vimentin+, troponin T+, troponin I+, α-actinin+, and connexin 43+. These findings highlight the importance of not only defining the acquisition of new markers (i.e., cardiac) but also characterizing in detail the expression of the markers defining the original cell type, which has not been done in previous studies assessing transdifferentiation of stem cells.

The mechanism for the partial transdifferentiation of MSCs toward a cardiac lineage is unclear. Several previous studies have suggested that cell fusion accounts for the observation of transdifferentiation [18, 23, 24, 46]. To determine whether our results are explained by cell fusion between rat embryonic cardiomyocytes and murine MSCs, we conducted coculture studies with male embryonic cardiomyocytes and female MSCs, using fluorescence in situ hybridization for X and Y chromosomes to identify fused cells. We demonstrated that although cell fusion does occur in this system (i.e., 0.68% fused cells), the phenomenon cannot account for all of the GFP+ MSCs that arise in the cocultures (6.3% GFP+). The frequency of fusion in our experiments is in agreement with other studies showing fusion levels of less than 1% among bone marrow progenitor cells [23, 46].

Thus, when cocultured with rCMs in vitro, adult bone marrow-derived MSCs demonstrate plasticity toward, rather than full transdifferentiation to, the cardiomyocyte lineage, as a result of altered genetic programming. This induction toward the cardiomyocyte lineage is presumably induced by local cell-to-cell contact and/or the presence of soluble factors in the coculture medium. For example, previous studies have shown that MSCs can establish connections with cardiomyocytes through gap junction channels composed of connexin-43 [21, 39, 47, [48], [49]50], a gene that we found was induced in our GFP+ MSCs. MSCs with gap junction connections to cardiomyocytes do not generate APs (in agreement with our findings) but can affect electrical conduction by electrotonic interactions, thereby either inducing [48] or preventing [47, 49, 50] arrhythmias.

Some limitations of our study need to be acknowledged. First, we evaluated the potential for MSC transdifferentiation in vitro. This was done so that we could definitively assess the phenotype of putatively transdifferentiated cells using single-cell patch-clamp electrophysiology and detailed immunophenotyping. The induction of cardiac gene expression and detection of cardiac markers by immunohistochemistry in our MSCs is very similar to what has been described for MSCs in vivo [12, 14, 15], suggesting that our model is relevant. Nevertheless, it remains possible that the plasticity of bone marrow-derived MSCs is different in vivo. For example, it may be possible that the in vivo setting provides unknown stimuli or a unique timing of exposure to growth factors that could increase the capacity for MSC transdifferentiation.

Other studies have suggested that MSCs can undergo true transdifferentiation on the basis of the demonstration of organized sarcomeres and the measurement of action potentials [11, 51, 52]. There are several possible explanations for these results. The study by Nishiyama et al. [51] examined umbilical cord blood-derived MSCs, whereas we studied bone marrow-derived MSCs, suggesting that different sources of MSCs could yield different patterns of results. A recent report from Pijnappels et al. [52] used neonatal bone marrow-derived MSCs to demonstrate electrical transdifferentiation, suggesting that the stage of development of the organism from which the MSCs are derived may be critical in determining the potential for multipotency of MSCs. It is important to note, however, that the neonatal MSC cultures studied by Pijnappels et al. [52] showed more than 10% CD45+ cells (compared with only 1% CD45+ cells in our cultures), suggesting contamination by hematopoietic stem cells. This makes their results difficult to interpret, as hematopoietic stem cells have been shown to undergo cardiomyocyte-like transdifferentiation [53]. Thus, our results suggest that the ability of adult bone marrow-derived MSCs to transdifferentiate into cardiomyocytes is limited in vitro; however, MSCs may yet be exploited for the generation of new cardiomyocytes by altering the source and/or treatment of the MSCs. For example, specific genetic modifications might increase the multipotent potential of MSCs, as has recently been demonstrated for other fibroblastic cells [54, 55].

Although it is possible that a species difference (mouse vs. rat) in our cocultures is a limiting factor or that mouse MSCs do not have the same capacity to transdifferentiate as do MSCs of larger mammals (i.e., humans), this seems unlikely. The MSCs in this system activated a cardiac gene program under coculture conditions that resulted in the expression of cardiac proteins similar to those shown in several other studies, suggesting that we have created a suitable microenvironment for studying transdifferentiation. Furthermore, several reports have shown that MSCs from one species can integrate with cardiomyocytes of other species [19, 31, 47, 56].

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

We have used a novel model to directly assess the ability of bone marrow-derived MSCs to acquire a cardiomyocyte phenotype under in vitro cell culture conditions. Our data show that although MSCs demonstrate cell lineage plasticity by acquiring cardiomyocyte markers, they do not develop the electrical properties of true cardiomyocytes under these specific conditions. The MSCs retain their stromal characteristics and hence do not undergo true transdifferentiation, suggesting that the plasticity of unmodified adult bone marrow-derived MSCs that have been expanded in culture is highly limited. On the basis of these data, it is conceivable that some of the improvements in cardiac function that have been described following MSC injection in the setting of heart disease are unrelated to the generation of new cardiomyocytes [1, 7, 8]. Further work will be required to investigate the mechanism(s) by which MSCs may improve cardiac function. Our observations may explain some of the controversies in the literature surrounding the ability of MSCs to undergo transdifferentiation to the cardiac lineage and highlight the importance of extensively characterizing the cells under investigation, especially with functional assays.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

We thank Steven Doyle, James Jonkman, Julie Ann Panakos, and Mingda Shi for expert technical assistance. A.K. holds the Gloria and Seymour Epstein Chair in Cell Therapy and Transplantation at the University of Toronto and University Health Network. This work was supported by funding from the Canadian Institutes for Health Research (MOP 79460) to P.H.B., who is a Career Investigator with the Heart and Stroke Foundation of Ontario. R.A.R. is the recipient of Post-Doctoral Fellowships from the Heart and Stroke Foundation of Canada, the Alberta Heritage Foundation for Medical Research, and the Canadian Institutes of Health Research-Tailored Advanced Collaborative Training in Cardiovascular Sciences program.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
SC-08-0329_Suppl_Fig_1.pdf215KSupplemental Figure 1
SC-08-0329_Suppl_Fig_2.pdf160KSupplemental Figure 2
SC-08-0329_Suppl_Fig_3.pdf250KSupplemental Figure 3
SC-08-0329_Suppl_Fig_4.pdf173KSupplemental Figure 4
SC-08-0329_Suppl_Fig_5.pdf162KSupplemental Figure 5
SC-08-0329_Suppl_Fig_6.pdf204KSupplemental Figure 6
SC-08-0329_Suppl_Table.tif33KSupplemental Table

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.