SEARCH

SEARCH BY CITATION

Keywords:

  • Cardiac;
  • Myocardium;
  • Transdifferentiation;
  • Stem cells;
  • Bone marrow;
  • Macrophages

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Since rates of cardiomyocyte generation in the embryo are much higher than within the adult, we explored whether the embryonic heart would serve as useful experimental system for examining the myocardial potential of adult stem cells. Previously, we reported that the long-term culturing of adult mouse bone marrow produced a cell population that was both highly enriched for macrophages and cardiac competent. In this study, the myocardial potential of this cell population was analyzed in greater detail using the embryonic chick heart as recipient tissue. Experiments involving the co-incubation of labeled bone marrow cells with embryonic heart tissue showed that bone marrow (BM) cells incorporated into the myocardium and immunostained for myocyte proteins. Reverse transcription-polymerase chain reaction analysis demonstrated that the heart tissue induced bone marrow cells to express the differentiated cardiomyocyte marker α-cardiac myosin heavy chain. The cardiomyocyte conversion of the bone marrow cells was verified by harvesting donor cells from mice that were genetically labeled with a myocardial-specific β-galactosidase reporter. Embryonic hearts exposed to the transgenic bone marrow in culture exhibited significant numbers of β-galactosidase-positive cells, indicating the presence of bone marrow-derived cells that had converted to a myocardial phenotype. Furthermore, when transgenic mouse BM cells were injected into living chick embryos, donor cells incorporated into the developing heart and exhibited a myocardial phenotype. Immunofluorescence analysis demonstrated that donor BM cells exhibiting myocyte markers contained only nuclei from mouse cells, indicating that differentiation and not cell fusion was the predominant mechanism for the acquisition of a myocyte phenotype. These data confirm that adult mouse bone marrow contain cells with the ability to form cardiomyocytes. In addition, the predominance of the macrophage phenotype within the donor bone marrow cell population suggests that transdifferentiation of immune response cells may play a role in cellular regeneration in the adult.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Evidence has been presented by several laboratories that stem cells exist in the adult bone marrow, possessing the capacity to generate new cardiomyocytes [19]. However, not every laboratory has been able to replicate these findings, and thus this topic has been controversial. Although some laboratories have reported great success in using donor adult stem cells to heal the damaged myocardium of mice who suffered an experimentally induced infarct [57], similar studies reported that no evidence could be found for the contribution of transplanted stem cells to forming new myocytes in the injured heart [1012]. One explanation given for this discrepancy is that the stem cells used in these studies simply do not have cardiac potential [13]. However, several other factors may account for the divergent experimental outcomes, because the procedures are complex, and significant experimental variations may exist among the different studies. For example, variables that may contribute to distinct experimental outcomes include the extent and localization of the induced infarct, variations in the stem cell populations used, and the precise protocol for delivering the stem cells into the wounded myocardium. Another possible variable is the processing of the hearts following the experimental interventions, because the use of frozen sections may not have fully preserved the damaged cardiac tissue and could account for negative assessments on the contributions of injected stem cells to the infarcted myocardium. Thus, the complexities of the procedures involved using infarcted adult heart tissue may constrict its use as an assay for directly assessing the myocardial potential of a cell population.

An environment that may be more conducive in permitting cardiac-competent cells to realize their myocardial potential is the embryonic heart. Since the rate of cardiomyocyte generation within the embryo is far greater than within the adult [14, 15], the embryonic heart probably provides a more supportive environment for both accepting and nurturing newly introduced cells. Thus, the embryonic heart may be a better vehicle than the adult heart for strictly testing the cardiac potential of adult bone marrow cells. In a previous study, we described a culture protocol for examining myocardial competence in which cells from adult mouse bone marrow were co-incubated with embryonic chick heart tissue [2]. Upon exposure to the embryonic heart explants, bone marrow cells displayed an ability to incorporate into the cardiac tissue and express muscle proteins. Although we observed evidence of cardiomyocyte formation from hematopoietic stem cells, surprisingly, the bone marrow cell population that appeared to exhibit the most consistent ability to transdifferentiate to a myocardial phenotype were macrophages. To follow up on that intriguing but preliminary observation, we have undertaken a more rigorous examination of the myocardial competence of adult mouse bone marrow macrophages, both by exposing these cells to embryonic heart tissue in culture and by direct microinjection into the hearts of living chick embryos. The results presented here conclusively show that cells from the adult bone marrow possess myocardial potential and demonstrate the utility of the embryonic chick as a vehicle for testing cell phenotypic potential.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Bone Marrow Isolation and Culture

Whole bone marrow was isolated from femurs of 8–12-week-old wild-type ICR, transgenic enhanced green fluorescent protein (EGFP) [16] (kindly provided by M. Okabe, Osaka University, Japan), or lacZ reporter TG1852 mice [17]. As described previously, the TG1852 mice were generated using an enhancer from the GATA-6 gene [17, 18] to drive the myocardial-specific expression of the lacZ reporter [17]. After flushing the bones with Iscove's modified Dulbecco's medium (IMDM; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), bone marrow was repetitively passaged through a 20-gauge needle and filtered through a 40 μm nylon sieve. The resulting cell suspension was washed and placed in bacterial grade Petri dishes with IMDM-20% fetal bovine serum (FBS) for four weeks, with fresh medium provided weekly. This incubation protocol highly enriches bone marrow cultures for macrophages, as previously described [19, 20].

Coculture of Embryonic Cardiac Tissue Explants and Bone Marrow

Chick eggs were incubated for 2–2.5 days to develop embryos corresponding to Hamburger and Hamiliton (HH) stage 16 [21]. Embryos were harvested, and following their excision, the hearts were briefly exposed to trypsin to produce tissue fragments. Isolated tissue was cultured in 8-well chamber slides in IMDM-20% FBS at approximately 25 tissue fragments per well and allowed to attach to the culture plastic. Thereafter, mouse bone marrow cells were added to the same wells at a concentration of 1 × 105 cells per well and incubated for 4 additional days. In selected experiments, carboxyfluorescein diacetate succinimidyl ester (CFDA; Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) was used to label bone marrow cells according to previously described protocols [2].

Microinjection and Culture of Embryos

Fertilized chicken eggs were allowed to develop in a 37°C egg incubator until HH stage 19–21. On the 3rd day, the eggs were cracked, and the entire contents including the albumen, yolk, and developing embryo carefully transferred to a plastic weigh boat, with the embryo oriented to lie on top. With the aid of a dissecting microscope, mouse donor cells were injected via a beveled-edged glass needle connected to a syringe, with microinjection volumes ranging from 0.5 to 10 μl. After injection, embryos were covered with a round 100-mm Petri dish and returned to the incubator for 2–3 more days.

Xgal and Immunofluorescent Staining

lacZ expression was determined by assaying for β-galactosidase activity, as visualized by Xgal staining. Explants and embryos were fixed in 4% paraformaldehyde and rinsed several times with 100 mM sodium phosphate (pH 7.3), 2 mM MgCl2, 0.01% sodium deoxycholate, 0.02% NP-40. Subsequent staining reactions with 5 mM potassium ferrocyanide and 1 mg/ml Xgal were closely monitored for color changes, which were stopped using 10% formalin. For the embryos, clearing of the tissue with methyl salicylate was required. Immunohistochemistry was performed using previously described protocols [2]. Cultures were methanol- or paraformaldehyde-fixed and then exposed to individual primary antibodies for specific molecular detection. The mouse monoclonal antibodies used to detect the nuclear marker QCPN, titin (9D10), and sarcomeric myosin heavy chain (MF20) were generated by Bruce Carlson, Jean Carlson, Marion Greaser, and Donald Fischman and obtained from the Developmental Studies Hybridoma Bank (developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA, http://www.uiowa.edu/∼dshbwww/). Tissue was also stained with mouse and rabbit antibodies purchased from Sigma that were specific for sarcomeric α-actinin and N-cadherin, respectively. For β-galactosidase detection, a protein-specific antibody (Rockland Immunochemicals, Inc, Gilbertsville, PA, http://www.rockland-inc.com) was added to live cells. Live cultures were washed once with sterile Dulbecco's phosphate-buffered saline (DPBS) following by a 30-minute incubation on ice with the primary antibody diluted 1:50 in DPBS with 10% dimethyl sulfoxide plus 10% goat serum. After two washes in DPBS, cultures were fixed with 4% buffered paraformaldehyde for 30 minutes at room temperature. Immunostaining for these various molecules was visualized following a 1 hour incubation at room temperature with the appropriate fluorescein- and rhodamine-labeled secondary antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) and several DPBS washes. Secondary antibodies used in this study were anti-mouse IgG (for QCPN, MF20, and α-actinin), anti-mouse IgM (for titin), and anti-rabbit IgG (for N-cadherin and β-galactosidase). After immunostaining, selected cultures were counterstained with 4′,6-diamidino-2-phenyindole (DAPI; In-vitrogen, Carlsbad, CA, http://www.invitrogen.com) to detect nuclei.

Reverse Transcription-Polymerase Chain Reaction Analysis

RNA was harvested from cultures and excised embryonic hearts using RNeasy kits (Qiagen, Valencia, CA, http://www1.qiagen.com). OneStep reverse transcription-polymerase chain reaction (RT-PCR) kits (Qiagen) were then used to amplify RNA template. Template concentrations were first normalized by 25 cycles of RT-PCR amplification for both mouse and chick homologs of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Thereafter, equal amounts of template were subjected to 30 amplification cycles with cardiac gene-specific primers, with each reaction run in parallel using the iCycler gradient thermal cycler (Bio-Rad, Hercules, CA, http://www.bio-rad.com). The hybridization temperatures for amplifying each of the cardiac genes were previously optimized for the individual primer sets. Primer sequences used in this study were mouse GAPDH (accacagtccatgccatcac + tccaccaccctgt-tgctgta), mouse α-cardiac myosin heavy chain (MyHC) (ggaagagtgagcggcgcatcaagg + ctgctggagaggttattcctcg), chick GAPDH (acgccatcactatcttccag + cagccttcactaccctcttg), and chick Nkx2.5 (ccttccccggcccctactac + ctgctgcttgaaccttctct). To verify the size of the amplified DNA fragments, PCRs were run in parallel with molecular weight markers on gels made from DNA agar (Marine BioProducts, BC, Canada, http://www.marbio.com).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Coculture of Bone Marrow Macrophages with Embryonic Heart Tissue

Macrophages were obtained from adult mouse bone marrow by selection of cells that are able to survive long-term incubation on hydrophobic, bacterial-grade plastic dishes. As reported previously, these cultures produced highly enriched populations of macrophages, as established by both functional assays and phenotypic analyses [2]. For example, in response to lipopolysaccharide stimulation, long-term bone marrow (LT-BM) cells exhibited high phagocytic activity, as measured by their incorporation of phagocytosable, fluorescent particles. As a comparison, the average number of particles ingested per LT-BM cell was 5-fold and 28-fold greater than observed with the positive control J774A.1 macrophage cell line and primary bone marrow cultures, respectively. The purity of the macrophage phenotype within these cultures was indicated by the presence of significant numbers of fluorescent particles within every cell contained in the LPS-treated LT-BM cultures. In addition, nontreated LT-BM cells showed positive reactivity for both Flt-1 and Mac-1 (CD11b) at the cell membrane [2], which is a phenotype characteristic of differentiated macrophages [22, 23]. Moreover, when these cells were placed in methylcellulose cultures and histochemically stained, they displayed the typical vacuolated phenotype of a macrophage (Fig. 1). No other cell phenotypes were observed under these conditions. Thus, by these criteria, macrophages appear to be the predominant cell type selected in the LT-BM cultures.

The myocardial potential of the adult mouse bone marrow macrophages was examined by their co-incubation with embryonic chick heart tissue. Hearts were removed from tubular stage chick embryos (HH stage 16) and briefly exposed to trypsin to produce small fragments of contractile tissue. Following the attachment of the cardiac tissue to the culture plastic, LT-BM cells were added to the wells, and the cocultures were incubated for 4 days prior to immunohistochemical analysis. To facilitate their identification within the cocultures, bone marrow cells were obtained from enhanced green fluorescent protein (EGFP)-expressing transgenic mice. Figures 2A and 2B show confocal microscopy optical z-series projections of contractile chick heart fragments within the cocultures, which have been stained for expression of either sarcomeric MyHC or sarcomeric α-actinin, respectively. As is typically observed, multiple BM cells will rest either on top of or circumferential to individual cardiac fragments. In addition, some of these cells will integrate into the cardiac tissue and appear to exhibit muscle protein expression. Better evidence for this latter assertion is provided by images of optically sectioned tissue, which show EGFP-labeled cells that had incorporated into the heart explants (Fig. 2C–2F). In these instances, the LT-BM cells that had integrated into myocardial tissue stained positive for sarcomeric α-actinin.

Another means to identify cell origin is to immunostain cultures for markers that are expressed by mouse but not chick tissue. Surprisingly, we found such a marker in the QCPN antibody, which has been well characterized previously for its ability to label cells of quail but not chick origin [24, 25]. While confirming the inability of this antibody to react to chick tissue, we discovered that QCPN is an excellent marker of mouse nuclei (Fig. 3A, 3B), and it thus allowed us to perform chick heart tissue coculture experiments with bone marrow from adult wild-type mice. Marking cultures with QCPN antibody allowed us to determine whether the donor mouse LT-BM cells that become associated with the explants were capable of fully integrating into the recipient cardiac tissue. To address this issue, heart/LT-BM cell cocultures were immunostained for N-cadherin, which is a cell junctional protein whose upregulated expression is associated with the cardiac tissue formation [26]. The shared N-cadherin positive cell borders of QCPN-stained mouse cells with neighboring cardiomyocytes indicated that donor cells had truly integrated into the contractile explants (Fig. 3C–3F).

Gene Expression Analysis of Cocultures

To verify the myocardial differentiation of mouse LT-BM cells that were co-incubated with heart tissue, RNA was isolated from the cocultures and assayed for species-specific gene expression (Fig. 4). For these experiments, parallel cultures were set up consisting of either adult mouse LT-BM cells only, embryonic chick heart tissue only, or both mouse LT-BM cells and chick heart tissue. Four days later, RNA was harvested from the cultures and PCR-amplified using species-specific oligonucleotide primers. As a further control, RNA was also harvested from embryonic day (ED) 16 mouse hearts. To normalize template concentration, we first performed separate amplification reactions to mouse or chick GAPDH housekeeping genes. Thus, concentrations of RNA template for both the ED16 mouse heart and mouse bone marrow only controls were normalized to volumes of the mouse bone marrow/chick heart coculture RNA that produced equal amounts of mouse GAPDH expression, whereas RNA concentrations of the chick heart-only cultures were matched to the amount producing equal amounts of chick GAPDH expression with the cocultures. Subsequently, all four RNA samples were subjected to RT-PCRs that independently amplified chick and mouse orthologs of cardiac-specific genes. Samples were amplified for the chick ortholog of the cardiac-transcription factor Nkx2.5 as a control for the presence of chick cardiac tissue. To detect the presence of fully differentiated cardiomyocytes of mouse origin, samples were assayed for expression of the mouse ortholog of α-cardiac MyHC. As shown in Figure 4, control amplification reactions produced the expected results. RNA isolated from cultures containing chick heart tissue and normalized to GAPDH exhibited similar levels of chick Nkx2.5 message, whereas neither chick-specific GAPDH nor Nkx2.5 sequences were detected in samples containing mouse RNA only. Accordingly, control mouse GAPDH primers prominently detected expression in the mouse heart, mouse bone marrow, and mouse bone marrow/chick heart coculture samples, but were unable to amplify sequence from chick heart only RNA. The control mouse and chick heart samples produced the expected positive and negative amplification with the mouse α-cardiac MyHC-specific primers. As anticipated, α-cardiac MyHC expression was not detected when mouse bone marrow cells were cultured by themselves. However, the mouse homolog of α-cardiac MyHC was readily amplified from the mouse bone marrow/chick heart cocultures. Thus, co-incubation with embryonic chick heart tissue provoked α-cardiac MyHC expression from adult mouse LT-BM cells, providing strong evidence that bone marrow cells will give rise to differentiated cardiomyocytes when exposed to a myocardial environment.

Bone Marrow from Transgenic Mice with Myocardial-Specific lacZ Expression

The initial experimentation presented above suggested that co-culture with embryonic heart explants provides a promising in vitro model for examining the capacity of adult bone marrow cells to integrate and differentiate into myocardial tissue. However, there is difficulty in characterizing unequivocally the phenotype of individual cells within a larger tissue using dual fluorescent analysis. Therefore, to determine definitively whether the bone marrow-derived cells actually assumed a cardiomyocyte phenotype, donor cells were isolated from transgenic mice that exhibit a myocardial-specific reporter gene. The donor mice used in the remainder of this study (referred to as TG1852 mice) were generated by inserting into the genome an Escherichia coli lacZ (i.e., β-galactosidase) transgene, whose expression is driven by a myocardial-specific enhancer from the GATA6 gene [17, 18]. Previous studies demonstrated that lacZ expression in these mice is totally restricted to the myocardial cell lineage [17]. In the developing TG1852 mouse embryo, lacZ expression is first turned on within the precardiac mesoderm and then displayed subsequently throughout the entire myocardium. For the present study, adult TG1852 mice were used as a source of bone marrow to establish cultures of LT-BM cells. When cultured by themselves, TG1852 LT-BM cells did not exhibit lacZ expression, as determined by absence of staining with the chromogenic substrate Xgal (Fig. 5A). As a further control, transgenic LT-BM cells were cocultured with noncardiac tissue—in this case, stromal cells from wild-type mice. Again, Xgal staining was not apparent (Fig. 5B, 5C). However, when LT-BM cells from the transgenic mice were co-incubated with embryonic chick cardiac tissue, many Xgal-stained cells were observed in the cultures. As shown in Figure 5D–5G, each contractile cardiac tissue fragment contained multiple cells that were lacZ-positive, indicating cells of adult mouse bone marrow origin that now exhibited a myocardial phenotype.

To further assess the differentiation of TG1852 LT-BM cells in the cardiac cocultures, cell phenotype was examined by fluorescent staining with antibodies specific for β-galactosidase. As shown above when assayed by Xgal staining (Fig. 5A), the absence of β-galactosidase immunoreactivity indicated that TG1852 LT-BM cells do not express the lacZ reporter gene when cultured by themselves (Fig. 6A, 6B). However, co-incubation of TG1852 LT-BM cells with embryonic cardiac explants did produce multiple β-galactosidase immunoreactive cells that were associated with each of the contractile heart tissue aggregates (Fig. 6C–6E). Some of the β-galactosidase-positive mouse cells were negative for sarcomeric protein expression (Fig. 6F), indicating that they were specified to the myocardial lineage but still possessed a precardiac phenotype. However, other cells were identified that dual stained for both β-galactosidase and sarcomeric proteins (Fig. 6G–6I), which demonstrates that cells marked with a myocardial lineage-specific reporter gene can undergo conversion to a differentiated myocyte phenotype.

Transdifferentiation of LT-BM Cells to Myocytes Without Cell Fusion

A common caveat offered to explain phenotypic plasticity is that the acquisition of phenotypic markers by a given cell may be due to its fusion with a differentiated cell [27, 28]. To address this concern, we changed the experimental design to determine whether we could prove definitively that adult bone marrow cells could acquire a myocardial phenotype by differentiation without cell fusion. As was done in the previous experimentation, cocultures of embryonic chick heart tissue and TG1852 mouse LT-BM cells were established and incubated for 4 days. However, following the incubation period, cultures were lightly exposed to trypsin, which produced suspensions of single cells and small cell clusters. The cell suspensions were replated at low density, allowed to reattach to the culture plastic for an additional 16 hours, and thereafter immunostained for both β-galactosidase and QCPN expression. In addition, these cultures were counterstained with DAPI, which labels all cell nuclei regardless of species origin. Figure 7A–7C shows a cellular field containing an individual β-galactosidase-positive cell possessing a single nucleus that is QCPN-positive. Figure 7D–7F shows a cell undergoing division that is highly immunoreactive for β-galactosidase, with both of its dividing nuclei exhibiting QCPN-immunoreactivity. Further evidence for the transdifferentiation of mouse LT-BM cells is indicated by expression of sarcomeric proteins by cells exhibiting individual QCPN nuclei (Fig. 7G–7I). In none of the experiments we performed was any evidence observed of a myocyte that exhibited both mouse and chick nuclei.

Microinjection of LT-BM Cells into the Developing Embryo

The experiments described above demonstrate that cells from adult mouse bone marrow can differentiate to cardiomyocytes in culture. As a next step, we extended this analysis by testing the capacity of adult mouse bone marrow cells to undergo myocardial differentiation within a living whole animal. Thus, LT-BM cells were injected into the hearts of tubular heart stage chick embryos and allowed to develop for an additional 48 hours before the embryos were processed for Xgal staining. As shown in Figure 8, donor mouse cells integrated into the chick myocardium and exhibited a cardiomyocyte phenotype, as indicated by lacZ expression. These data demonstrate that cells obtained from adult mouse bone marrow are capable of undergoing cardiac differentiation in situ. Taken together, these data confirm that cells within the adult bone marrow possess myocardial potential and have the capacity to generate new cells for the myocardium.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

A subject that has been mired in controversy concerns whether adult stem cells have utility for generating new myocytes in the diseased or damaged adult heart. This issue has been principally investigated by examining the ability of stem cells to repair experimentally induced myocardial infarcts within adult mice. Although some laboratories have reported success in using stem cells from the bone marrow to regenerate myocytes within a damaged myocardium [57], other laboratories have been unable to replicate these findings [1012]. The negative results have been cited as evidence to support the assertion that bone marrow cells do not in fact have myocardial potential [13]. However, many other possible explanations could account for such a dramatic discrepancy in the experimental outcomes, because the healing of a severely damaged myocardium from an induced infarct is a complex undertaking. Moreover, alterations in the extracellular matrix within the infarct zone may result in conditions that are far from optimal for generating new myocytes [29, 30]. Perhaps the infarcted adult heart assay, although highly pertinent to future treatments of adult cardiac disease, is not necessarily the best experimental system for directly testing the myocardial potential of a cell population. A healthy adult heart would also not serve as a choice setting for assessing the cardiac competence of injected donor cells, because the rate of cardiomyocyte regeneration in the adult is normally very low. In contrast, changes in tissue composition are more dynamic within the developing embryo than within the adult. During embryonic development, the heart not only undergoes substantial growth, but also incorporates various cellular populations as constituents of the myocardium [14, 15]. For these reasons, we considered the embryonic environment as an alternative experimental system for determining the phenotypic capacities of adult donor cells.

The Embryonic Chick as a Model for Examining Adult Stem Cell Potential

The present study demonstrates that the embryonic chick is an effective vehicle for studying the cardiac competence of adult mouse bone marrow cells. Several attributes make the chick embryo a practical experimental model to study myocardial cell potential. The cellular and structural composition of the mammalian and avian heart is highly similar [31]. Precisely staged embryos are much more easily obtained from bird eggs than they are from the mammalian uterus. Cardiac tissue can be readily harvested from the chick embryo for explant studies, which serve as a screen for testing the myocardial potential of donor cells within an in situ-like environment. Donor cells can be microinjected into living chick embryos with minimal difficulty. Microinjected embryos can subsequently develop normally within a resealed eggshell or cultured in a plastic vessel. Thus, the living chick embryo is particularly advantageous as a recipient for cell grafting studies, because the experimentation (including the microinjections, the survival rates of the recipients, and subsequent data analysis) is easier and much less likely to produce complications than would grafting directly into adult mammalian hearts.

The experiments described here employed the embryonic chick as both a source of tissue explants for culture experimentation and as a whole animal model system. When prelabeled adult mouse bone marrow cells were co-incubated with embryonic chick heart tissue, some of the bone marrow cells integrated into the cardiac tissue and adopted a cardiomyocyte phenotype, as indicated by the dual fluorescence of immunostained sarcomeric proteins and donor tissue-specific label. The presence of bone marrow cell-derived cardiomyocytes was verified by harvesting cocultures following the incubation period and replating at low cell density, which allowed for the unambiguous identification of multifluorescent cells. The myocyte phenotype of mouse bone marrow cells within the cocultures was further substantiated by gene expression analysis using RT-PCR primers specific for the mouse ortholog of α-cardiac myosin. Myocardial differentiation was also observed when mouse bone marrow cells were directly injected into the hearts of developing chick embryos, which confirmed the myocardial competence of the donor cells using a whole animal model. Moreover, the ability to use the QCPN antibody to distinguish mouse from chick nuclei enabled us to confirm that mouse bone marrow-derived cells are able to acquire a myocardial phenotype by transdifferentiation. These data demonstrate that the embryonic heart is a supportive environment for both accepting and nurturing newly introduced cells and an excellent vehicle for testing stem cell potential. The utility of this model was further augmented using donor cells carrying a cardiomyocyte lineage-specific molecular tag.

Tagging Bone Marrow Cells with a Myocardial-Restricted Transgene Reporter

When stem cells are introduced into a tissue, either in culture or within an animal, a method for marking the cellular graft must be employed. For example, stem cells can be prelabeled with fluorescent dye or obtained from transgenic animals that universally and constitutively express a reporter gene, and then placed within wild-type tissue. After a period of incubation, changes in the phenotype of the transplanted stem cells can be assessed by dual fluorescence analysis for the label on the grafted cells and an immunostained differentiation marker. The problem associated with this experimental approach is that the unambiguous identification of the transplanted, labeled cells within a tissue is not always possible, even when the tissue is optically sectioned by confocal microscopy or physically sectioned. A common criticism of these types of studies is that what may appear to be a single cell exhibiting dual fluorescence is actually a labeled, nondifferentiated donor cell superimposed above an immunostained host cell. The way to lessen the experimental ambiguity is to harvest donor cells from transgenic animals that only exhibit a differentiation-specific label. Thus, for studies examining the myocardial potential of adult bone marrow cells, we used as a source of donor cells TG1852 transgenic mice that express a lacZ reporter gene exclusively in cells that are undergoing myocardial differentiation [17].

Obtaining bone marrow from TG1852 mice allowed for the unambiguous identification of donor cells that had undergone myocardial differentiation within cardiac tissue. This phenotypic conversion of bone marrow cells was observed both within chick cardiac explant cocultures and upon the direct delivery of donor mouse cells within the hearts of living chick embryos. In the whole embryo, lacZ-expressing donor bone marrow cells were clearly observed within the outer wall of the developing heart. Within chick heart cocultures, the myocardial-specific lacZ tag detected greater numbers of positive mouse donor bone marrow cells than did sarcomeric protein immunoreactivity. That this result was not due to leakiness of the lacZ marker was indicated by the lack of both Xgal staining and β-galactosidase immunoreactivity when bone marrow cells were cultured in the absence of cardiac tissue. Instead, it is likely that this finding is reflective of the lacZ expression in TG1852 embryos, which marks both differentiated cardiomyocytes and myocardial-committed cells within the precardiac mesoderm [17]. Thus, this lacZ marker is capable of identifying cells in the earliest stages myocardial differentiation and indicates that the myocardial competence of cells within the bone marrow may be significantly greater than what is apparent by only characterizing differentiated myocyte phenotypes. It should be noted that the assays used in this study were not intended to optimize cardiac conversion but were designed primarily to provide an unambiguous visual demonstration of BM cell transdifferentiation to cardiomyocytes. Future experimentation that accurately quantifies the frequency of these events will allow for a more informed evaluation of the biological importance of transdifferentiation.

Macrophage Transdifferentiation to Cardiomyocytes

A principal conclusion of this study is that there are cells within the adult mammalian bone marrow capable of integrating into the myocardium and forming cardiomyocytes. Based on the experimental results presented here, this conclusion appears valid. However, our data have, in addition, led us to a further postulate. The donor cells used in these studies were harvested from long-term cultures of adult mouse bone marrow grown on bacterial grade Petri dishes. Based on both phenotypic and functional analyses, the predominating cell type within the long-term bone marrow cultures were macrophages [2]. Moreover, the ability to survive long-term incubation on hydrophobic, bacterial-grade plastic dishes has been described as a unique attribute of macrophages [19, 20]. Admittedly, it is possible that a contaminating stem cell within these cultures is the cardiac-competent population. However, given the macrophage predominance, the invoking of a distinct sparsely present myocardial progenitor within the cultures would necessitate a stem cell population exhibiting unprecedented rates of myocardial conversion. As an alternative, we suggest that the cardiomyocytes arising from LT-BM cultures were obtained via the transdifferentiation of macrophages. This is not to imply that macrophages are the only bone marrow cells possessing myocardial competence, because multiple cell types may have the ability to form cardiomyocytes. Instead, we believe our data suggest that the macrophage is a cell possessing an appreciable myocardial potential.

Why should macrophages possess myocardial potential? Previously, we proposed that the plasticity of cell potential is related to how new cell phenotypes were generated throughout evolution [2, 3, 32]. As more complex organisms evolved, more distinct types of specialized cells were spawned. Although the generation of new specialized cell types may have been the result of selective mechanisms acting on stem cells, the emergence of new differentiated cell types by variation from existing differentiated phenotypes would appear to be a more direct and less complicated mechanism for promoting evolutionary diversification.

In the adult, the greatest amount of cellular regeneration occurs during the wound healing response [33]. Thus, immune response cells, which normally are recruited to damaged tissue [34], may themselves be a contributing cellular source for repairing the wound site. Immune response cells also play a role in normal tissue homeostasis by their removal of dead and diseased cells [35]. The macrophage is the earliest evolved immune response cell, and it appears in many primitive organisms [36, 37]. Accordingly, we speculate that new cell phenotypes arose during evolution as variations of the macrophage, which may account for the ability of this cell to transdifferentiate into other specialized cell types.

thumbnail image

Figure Figure 1.. Long-term bone marrow (LT-BM) cells exhibit a macrophage phenotype. LT-BM cells were harvested, placed in methylcellulose cultures, and stained with May-Grünwald-Giemsa (MGG) dye. Low-magnification (A) and high-magnification (B) views of MGG-stained methylcellulose cultures of LT-BM cells demonstrate a phenotypically pure population of cells that exhibit the gray cytoplasm, purple nucleus, and vacuolated morphology typical of a differentiated macrophage. Scale bars = 25 μm.

Download figure to PowerPoint

thumbnail image

Figure Figure 2.. Long-term bone marrow (LT-BM) cells incorporate into cardiac tissue and express cardiac proteins. LT-BM cells obtained from enhanced green fluorescent protein (EGFP)-expressing transgenic mice were co-incubated with embryonic chick heart tissue for 4 days. Cultures were stained for sarcomeric MyHC (A) or α-actinin (B–F) (red) and imaged by confocal microscopy. (A): Cocultured tissue displayed as a z-series projection showing the association of multiple EGFP-labeled LT-BM cells with the heart tissue. (B): Low-magnification view of an α-actinin-immunostained heart fragment displaying a cluster of EGFP-positive cells (arrow), which shown at high magnification (C) in optical section appears to exhibit cells displaying dual EGFP and α-actinin fluorescence (arrows). (D): An optical section of a cocultured heart explant showing an EGFP-marked adult mouse bone marrow cell (arrow) that has integrated into the myocardium and exhibits α-actinin expression. (E, F): Optically sectioned contractile cardiac tissue fragment, which contains an EGFP-positive mouse donor cell (arrow). For this explant, the optical section is exhibited for α-actinin staining only (E) or both α-actinin and EGFP labeling (F). The juxtaposition of (E) and (F) demonstrates that the LT-BM cell exhibits positive reactivity for the sarcomeric protein. Scale bars = 20 μm.

Download figure to PowerPoint

thumbnail image

Figure Figure 3.. Integration of long-term bone marrow (LT-BM) cells into cardiac tissue. LT-BM cells obtained from wild-type mice were incubated in the absence (A, B) or presence (C–F) of embryonic chick heart tissue. (A, B): Phase and fluorescent image of a mouse LT-BM culture stained with the QCPN antibody (red). Note that this antibody, which does not react to chick cells, labels nuclei of all mouse cells. (B–F): Confocal microscopy images of cocultured mouse LT-BM cells and chick cardiac tissue, which was immunostained for both N-cadherin (green) and QCPN (red). (B): Low magnification view of a contractile cardiac tissue fragment showing prominent cell membrane expression of N-cadherin. The immunostaining with QCPN antibody reveals that multiple cells of mouse origin have become incorporated into the heart tissue. (D–F): Higher magnification views of the boxed areas in (C) demonstrate that QCPN-reactive mouse cells (arrows) share N-cadherin-positive cell borders with neighboring chick cardiomyocytes, indicating that LT-BM-derived cells have integrated into the cardiac tissue. Scale bars = 20 μm.

Download figure to PowerPoint

thumbnail image

Figure Figure 4.. Myocardial gene expression of mouse long-term bone marrow (LT-BM) and chick heart cocultures. From left to right are shown reverse transcription-polymerase chain reaction-amplified RNA from embryonic day 16 mouse heart and 4-day cultures of either adult mouse LT-BM cells, embryonic chick heart, or both chick heart and adult mouse LT-BM cells. Amplification reactions used either mouse-specific primers for α-cardiac MyHC and GAPDH or chick-specific primers for Nkx2.5 and GAPDH. Abbreviations: cHt, embryonic chick heart; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; mBM, adult mouse long-term bone marrow cells; mHt, mouse heart; MyHC, myosin heavy chain.

Download figure to PowerPoint

thumbnail image

Figure Figure 5.. Expression of myocardial-specific genetic reporter by mouse bone marrow cells. Long-term bone marrow (LT-BM) cells obtained from adult TG1852 mice were cultured for 4 days and Xgal-stained (blue) to reveal expression of myocardial-specific lacZ expression. (A): Mouse LT-BM cells cultured by themselves were Xgal-negative, because they exhibited <1 blue-stained cell per 2 × 105 cells. (B, C): Phase and fluorescent view of TG1852 LT-BM cells prelabeled with the green vital dye carboxyfluorescein diacetate succinimidyl ester and cocultured with stromal tissue from wild-type mice. Note that LT-BM cells (revealed by the fluorescent dye) remained Xgal-negative when cultured with noncardiac tissue. (D–G): Several examples of mouse LT-BM cell/embryonic chick heart cocultures. Note the presence of multiple blue Xgal-stained cells, which indicates cells of mouse bone marrow origin that exhibit a myocardial phenotype. The numbers of cells in the cocultures that stained with Xgal was >1 per 104 total LT-BM cells. Scale bars = 60 μm.

Download figure to PowerPoint

thumbnail image

Figure Figure 6.. Co-expression of myocardial-specific reporter and sarcomeric proteins by mouse bone marrow cells. Long-term bone marrow (LT-BM) cells obtained from TG1852 mice were incubated either by themselves and stained with β-galactosidase-specific antibody (green) (A, B) or in the presence of embryonic chick heart tissue and dual-immunostained for β-galactosidase (green) and the sarcomeric protein titin (red) (C–I). (A, B): Phase and fluorescent image of a mouse LT-BM culture demonstrates that these cells do not normally exhibit β-galactosidase immunoreactivity. (C–E): After 4 days of coculture, cardiac tissue fragments exhibited prominent titin-positive myofibrils and displayed multiple β-galactosidase immunoreactive cells, which are cells of mouse origin that have transdifferentiated to the myocardial lineage. Examples of these latter cells are shown in subsequent panels, which are high-magnification views of the areas highlighted by the boxes in (E). (F): A β-galactosidase-positive mouse cell (arrow) that did not exhibit titin staining, indicating a cell specified to the myocardial lineage but still possessing a precardiac phenotype. (G): Optical sectioning also demonstrated the presence of mouse donor cells with more differentiated cardiomyocyte phenotypes, as indicated by the display of a striated titin expression pattern by the β-galactosidase-positive cell (arrow). (H, I): Another area of the same cardiac tissue fragment exhibited in optical section, shown for titin reactivity only or both titin and β-galactosidase staining, respectively. The juxtaposition of these two latter panels demonstrates that the LT-BM cell (arrow) has converted to a differentiated cardiomyocyte phenotype and has integrated into the cardiac tissue. Scale bars = 40 μm.

Download figure to PowerPoint

thumbnail image

Figure Figure 7.. Display of a myocardial phenotype by bone marrow cells does not involve cell fusion. After 4-day incubations of TG1852 mouse long-term bone marrow (LT-BM) cells with chick cardiac tissue, the cultures were exposed briefly to trypsin and replated. Following an additional overnight incubation, the cultures were fluorescently stained for β-galactosidase (green), 4′,6-diamidino-2-phenyindole (DAPI) (blue), and QCPN (red) (A–F) or titin (green), DAPI (blue), and QCPN (red) (G–I). (A): A cellular field shown in phase contrast, with the corresponding β-galactosidase plus DAPI staining (B) and QCPN immunofluorescence (C). Arrows indicate an individual β-galactosidase-positive cell with a single nucleus that is QCPN-reactive. Note that this field also contains cells that only stained with DAPI, which are cells of chick origin. (D): Another cellular field shown in phase contrast, with the corresponding β-galactosidase plus DAPI staining (E) and QCPN immunofluorescence (F). Arrows indicate an individual β-galactosidase-positive cell undergoing cell replication, with both separating nuclei being QCPN-reactive. (G): A small cluster of cardiac tissue displayed in phase contrast, with the corresponding titin plus DAPI staining (H) and QCPN immunofluorescence (I). Arrows indicate an individual titin-positive cell with a single nuclei that is QCPN-reactive. Note that the titin-positive cardiac tissue fragment contains multiple cells that label with QCPN. These images, documenting the presence of cells whose nuclei are strictly of mouse origin and express either a myocardial-specific β-galactosidase reporter or sarcomeric proteins, demonstrate that cell fusion could not account for this acquisition of a cardiomyocyte phenotype. Scale bars = 20 μm.

Download figure to PowerPoint

thumbnail image

Figure Figure 8.. Myocardial differentiation of mouse bone marrow cells in situ. Long-term bone marrow (LT-BM) cells obtained from adult TG1852 mice were injected into the hearts of living chick embryos. After 2 days of further development, embryos were fixed and stained for Xgal. (A): Whole embryo view showing the heart (indicated by the box) with several blue cells. (B, C): Higher-magnification views of the heart from the same embryo showing blue cells in the myocardial wall. Xgal-positive cells are adult mouse bone marrow cells that have shifted to a cardiomyocyte phenotype after exposure to the myocardium. Scale bars = 90 μm.

Download figure to PowerPoint

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank Lisa Martin and Marlisa Sooy for expert technical assistance and Dr. Steve Kubalak for generous help. This work was supported by grants from the American Heart Association (0555522U), National Aeronautics and Space Administration Experimental Program to Stimulate Competitive Research (NCC5-5775), NIH IDeA Networks of Biomedical Research Excellence (2P20RR016461-05A1), and NIH National Heart, Lung, and Blood Institute (P01-HD39946 R01-HL55373, and R01-HL073190).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
SC050128SuppFile1.pdf99KSupplemental Figure Legends
SC050128SuppFile2.pdf141KSupplemental Figure 1
SC050128SuppFile3.pdf101KSupplemental Figure 2
SC050128SuppFile4.pdf71KSupplemental Figure 3

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.